National Academies Press: OpenBook

Advanced Ground Vehicle Technologies for Airside Operations (2020)

Chapter: Chapter 3 - Applications and Lessons Learned

« Previous: Chapter 2 - Background
Page 9
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 9
Page 10
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 10
Page 11
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 11
Page 12
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 12
Page 13
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 13
Page 14
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 14
Page 15
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 15
Page 16
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 16
Page 17
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 17
Page 18
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 18
Page 19
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 19
Page 20
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 20
Page 21
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 21
Page 22
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 22
Page 23
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 23
Page 24
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 24
Page 25
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 25
Page 26
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 26
Page 27
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 27
Page 28
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 28
Page 29
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 29
Page 30
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 30
Page 31
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 31
Page 32
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 32
Page 33
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 33
Page 34
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 34
Page 35
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 35
Page 36
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 36
Page 37
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 37
Page 38
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 38
Page 39
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 39
Page 40
Suggested Citation:"Chapter 3 - Applications and Lessons Learned." National Academies of Sciences, Engineering, and Medicine. 2020. Advanced Ground Vehicle Technologies for Airside Operations. Washington, DC: The National Academies Press. doi: 10.17226/26017.
×
Page 40

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

9 The most prevalent applications for AGVT have been in the roadway sector, with numerous companies developing and testing AGVT for passenger cars. The Google self-driving car project was announced in 2009 and by 2014 it had evolved to encompass a completely auto- nomous prototype vehicle without a steering wheel, gas pedal, or brake pedal (Hartmans, 2016). By that time, many major automotive companies were also developing technologies as well as partnering with software and tech companies to support automation in their vehicles. Although there is substantial publicity for milestone events, in many cases, development is proprietary, which makes it difficult to accurately predict the timeframe for deployment of advanced technologies. Other applications include construction and mining, ports, agriculture, and manufacturing. Documentation of applications in the aviation sector is limited; however, a variety of applications are discussed below, including snowplows, automated shuttles, jet bridge, and robotic tug applications. Related aviation applications such as aircraft parking guidance systems and ADAS systems for ground vehicles are also included in this discussion. Finally, a brief discussion of lessons learned is presented, including lessons learned from unmanned aerial systems (UAS) at airports, AV in the roadway sector, previous technology deployments at airports, and other considerations. Applications in Roadway Construction and Mining The construction environment is similar to the airport environment in a lot of ways. Both have a well-defined and constrained operating area with a lot of activity, both are owned by a single entity (but require coordination with numerous stakeholders), and both have trained personnel working on-site with limited public access. Given these similarities, it is worthwhile to investigate the applications of advanced technologies in construction. Construction Automation has been used in construction to increase equipment accuracy. One example is roadway milling and resurfacing, where Automated Machine Guidance (AMG) has been used on many projects, including I-70 in Colorado (Townes, 2014). AMG uses a 3D model to guide construction equipment more accurately and quickly, and leverages data provided by a GPS to determine the horizontal and vertical position of the equipment relative to a computer gener- ated model of the project. The complexity and ability of these systems have rapidly advanced to the point where they need minimal input from an operator. The I-70 project owner, Colorado DOT, used the traditional design-bid-build procurement process, of which the bid with AMG was the lowest qualified bid, and competed with bids using traditional processes. This demonstrates that AMG can be both efficient and economical for construction projects (Townes, 2014). C H A P T E R 3 Applications and Lessons Learned

10 Advanced Ground Vehicle Technologies for Airside Operations A Caterpillar study confirmed the benefits of using advanced technologies in a study that compared traditional and technological assisted machines on a 400-foot stretch of road in East Peoria, Illinois. To replicate an actual job site, they used a mix of different equipment manu- factured by both Caterpillar and their competitors. The results showed a clear advantage when using the technology compared to the traditional way of constructing a road (Caterpillar, n.d.): • 31% fewer man hours, • 34% fewer equipment hours, • 46% fewer project hours, and • 37% less fuel consumption. The obvious correlation between construction automation and airports is runway construc- tion. Similar equipment is used for paving in both roadway and runway construction, although there are more miles of roadway paved each year. Runway paving applications are discussed later, and it is valuable to note that the benefits have been demonstrated in both sectors. Mining Mining is a subsector of construction, in which automation has been successfully utilized to increase safety and productivity. Mining operations are also similar to airport operations, as they both have a limited operating area with well-defined geographical constraints. Auto- mated mining equipment includes both new equipment with automation, as well as the conversion of existing equipment into driverless technologies using software and hardware. Mining equipment that has been fully automated includes haul trucks, articulated dump trucks, bulldozers, and excavators (ASI, n.d.). Control software enables operators to move a variety of mining vehicles simultaneously and with improved accuracy. The entire fleet can be integrated using the same software and controlled remotely from a command center. Operators can assign routes, issue load and dump commands, and lock out areas of a map. The vehicle will automatically provide obstacle avoidance and real-time vehicle diagnostics. Different levels of vehicle-use permissions can be given based on job needs. Automation features increase safety and minimize human errors in vehicle opera- tions. Automation enables automated steering (aka guided steering) along a predetermined path; this reduces the need for high-resolution video or operator skills. The operator may control acceleration and braking while the software system controls the steering of the vehicle. This could be helpful in aviation to decrease the likelihood of runway incursions by defining vehicle paths that do not overlap with the movement area or runway safety area (depending on the application). The use of predetermined paths may also be helpful for snowplows, especially when visibility is limited. Fleet management capabilities allow operators to see the status and location of every piece of equipment on the work site in real time. Condition reporting, including slick or rough sections of road, can be sent from one vehicle to all others in the system. With all the vehicles inter- connected and communicating, vital information is constantly exchanged and evaluated. Lower levels of autonomy can also be incorporated into a fleet, including operator assist components which track the operator to combat fatigue and collisions, and also provide training when it detects best practices are not being used. Fatigue from operating heavy machinery for long hours is a hazard in both mining and snow removal operations at airports. Automated systems can monitor driver actions and send a warning if there is a trend of deviations beyond accepted driving norms. Collision warning systems in the vehicle also alert the driver to impeding collisions. Implementing these technologies into snow removal and other airside equipment could potentially help streamline operations and increase safety.

Applications and Lessons Learned 11 Applications at Ports Automation is successful at ports internationally and is currently being tested at ports in the United States. Although only four U.S. seaports use extensive automation technologies, the operational benefits of automation are immense. Automation of both cranes and trucks contributes to operational efficiencies, and ports are an ideal environment for implementing automated technologies because they are controlled, consistent, and require the same actions to be repeated continuously. The Ports of Long Beach and Los Angeles have many terminals and are the top two ports in the United States based on container volume, which is measured in twenty-foot equivalent units (TEUs) (World Shipping Council, 2016). Most of the cranes in these ports are driven manually, but a couple of terminals, notably the TraPac Terminal, have begun to implement advanced technologies and automation to increase efficiency. After enormous cranes hoist the container boxes up from their ships and deposit them dockside, automated cargo-haulers grab, lift, and place them in their next destination, which may be up to four stories high. TEUs are stacked by an autonomous stacking crane, and lifted again when it is time to be moved to a truck. There is no human operator necessary for this part of the operation, since humans are not able to locate and move the boxes as efficiently as AI. The new technologies being experimented with at TraPac terminal are considered the most efficient way to handle the steadily increasing amount of cargo being shipped across the globe. Automation is expected to increase seaport productivity by 30%. APM Terminals, part of Dutch shipping giant A.P. Moeller-Maersk, have been implementing automation since the 1990s and reports that their automated terminal in Rotterdam requires half of the human labor force that would be required if automation were not used (Dillow and Rainwater, 2018). GE has also partnered with the Port of Los Angeles to pilot the Port Information Portal, a program designed to digitize maritime data and create a dashboard that reflects the health and status of the port supply chain. Initiated with a short runtime and a small percentage of the overall operation, the program has been extended for 3 years and expanded to encompass the entire port (Dillow and Rainwater, 2018). On the ground, the Chinese company TuSimple has been making great strides towards fully automated truck operations at ports (shown in Figure 2), which illustrates an autonomous truck transporting cargo boxes in China during a trial run. Their fully autonomous trucks are able to safely drive on the highway without intervention and slowly drive around the port environment to locate cargo boxes. TuSimple utilizes an array of cameras and LiDAR for their operation, but reports that cameras are the better technology for the sensing needed for auto- mated trucks as they have a greater range and are less expensive than LiDAR (Zurschmeide, 2018). Figure 2. Automated truck operation at ports in China. Photo: TuSimple, 2018.

12 Advanced Ground Vehicle Technologies for Airside Operations Automated trucks for port applications are only implemented in China. TuSimple is developing test fleets for commercial use on the highways and local streets in the United States (S. Morris, September 28, 2019, personal communication; Cassidy, 2012). Challenges ports face when it comes to automated technologies include high upfront costs, opposition from labor unions, and the fact that there is no “one-size-fits-all” solution for auto- mating a port. There are many hardware and software components required for automation, and implementation at a port can cost more than $2 million an acre (Dillow and Rainwater, 2018). This is significant, given the fact that the Port of Los Angeles complex is 7,500 acres. Labor unions have also stepped up their resistance to automation; a recent 6-year labor union contract with the International Longshoremen’s Association prohibits fully automated terminal ports until 2024 (Morley, 2018). Each port is different and requires a unique solution, and the best automated solution for one port will not be the best solution for another, which is analogous to airports. However, there are similarities among ports such as the need for cranes to load and unload cargo containers, which must be organized in a logical and efficient manner to be easily located when it is time for them to move on. This task is ideal for automation, because it uses algorithms and massive data for maximum efficiency. Applications in Agriculture Agriculture is an excellent application for advanced technology and automation in the United States, where hundreds of acres of land have to be tilled, plowed, seeded, watered, and harvested several times a year. This requires hours of work, covering the exact same stretches of land. This repetition, as well as advances in data collection, provide an ideal environment for the application of AGVT. Control software allows farmers to create maps of fields, which are then used to support software that remotely controls autonomous tractors that follow a precision in-field guide path. Sensors in the vehicle can detect obstacles in the path using LiDAR and infrared tech- nologies and alert the farmer of their presence. The farmer can then approve the machine to follow an alternate pathway around the obstacle, or in some cases, drive through the obstacle (ASI, n.d.). Autonomous farming equipment helps address labor shortages in the farming industry, which is valuable as 71% of labor-intensive crop growers experience labor shortages, which prevent one in five farmers from finishing their harvest (California Farm Bureau Federation, 2012). Advanced technologies allow multiple units to be controlled from the same centralized plat- form, 24 hours a day. This allows machines to operate overnight, when labor is hard to schedule and when darkness may reduce productivity. Autonomous solutions can work in tandem with traditional farming equipment and tractors can switch between autonomous and manual opera- tion (ASI, 2016). Another option is to have a single operator drive a tractor platoon, with the autonomous vehicles following behind. Weather has a significant impact on farming operations, and the sensors and software are designed to protect the vehicle from harsh weather conditions. If deteriorating weather conditions are detected, the operator will be alerted and operations will cease until a safe oper- ating environment is restored. This can save money by ensuring that operations are limited during less-than-ideal conditions. Mowing applications that have been proven in agricultural may have direct applications for airside activities. Many airports have hundreds, or even thousands, of acres of grass that

Applications and Lessons Learned 13 need to be mowed regularly. Ensuring that grass is mowed to the appropriate height helps reduce wildlife hazards on the airfield, and proper upkeep is required by Part 139, Airport Certi- fication. Typically, these operations are handled by maintenance personnel, but automated mowing would free up maintenance personnel for other important tasks. The Kubota Tractor Corporation is currently working with Indianapolis-based Smart Guided Systems, in developing auto-steering tractors with the capability of autonomous and remote driving. The tractors are currently undergoing testing at many locations nationwide such as the Indianapolis International Airport and New Braunfels Regional Airport. Steering paths are accurate up to 2 centimeters and use pre-planned GPS routes and Android tablet control (Hill, 2017). Mowing applications are of significant commercial interest due to large markets that include golf courses. Applications in Manufacturing and Industry A study conducted by McKinsey found that as of 2015, 478 billion of the 749 billion working hours (64%) spent on manufacturing related tasks around the world could be conducted autonomously with established technology, which would allow $2.7 trillion in labor costs to be eliminated or repurposed. This puts the manufacturing industry in second place (after the food services industry) for automation potential (Chui et al., 2017.) Applications of advanced technology in manufacturing and industry include robots, automated carts, and automated processes. Robots Amazon began using robots in 2014 and is at the forefront of integrating robotics into ware- house operations. Amazon uses machines developed and produced by Kiva Systems, a robotics company Amazon bought for $775 million in March of 2012 and renamed Amazon Robotics (Wingfield, 2017). They now have over 100,000 robots in warehouses around the world, with plans to add more to that fleet. There are two categories of robots used within Amazon warehouses. Transport robots move products from storage areas in the warehouse to other locations for packing and shipment. Transport robots are able to move up to 3,000 pounds of product at a time and can move very close to each other without colliding (Berger, 2016). This capability decreases the square footage required for safe operation within the warehouse. The second type of robot is the robotic palletizer, with long yellow arms that pick up standard sized bins or pallets to move them along the supply chain (Wingfield, 2017). They have a large capacity and are able to lift heavy loads to increase efficiency (Lisota, 2017). Robots have allowed Amazon to increase autonomy and efficiency in the warehouse, although Amazon has also continued to increase the staff of human workers to keep up with growing demand. Tasks such as physical handling of heavy boxes have shifted to robots and human workers oversee the operation of the autonomous robots. This has led to less physically demanding and more mentally stimulating work for the employees and reportedly results in greater job satisfaction (Wingfield, 2017). Autonomous Carts Another AGVT company, Canvas, is looking to bring autonomous solutions similar to the Amazon fleet to companies around the world. They have developed the Canvas Autonomous Cart, which is able to autonomously navigate the congested and unpredictable spaces in a ware- house to transport products and supplies (Canvas Technology, 2017). Moving heavy materials

14 Advanced Ground Vehicle Technologies for Airside Operations from one place to another is a time-consuming and cumbersome task for employees in ware- houses, and Canvas seeks to eliminate that task altogether with their automated carts, which can replace 20 miles of walking per shift on a single charge (Canvas Technology, n.d.). One of the goals of Canvas is to create durable carts that can operate indoors and outdoors. The advanced 3D cameras embedded in the cart allow continual scanning of the environment and eliminate the need for the creation of customized maps. Canvas Carts also communicate with each other, sharing information that has been collected and calibrating their movement to the task and environment instantaneously (Canvas Technology, 2017). The carts are controlled by a web application that allows operators to set up stopping points and routes within the workplace environment. Once a cart is on the way to its destination, it will yield to obstacles or find an alternate route; this increases safety. The pressure to meet faster delivery time expecta- tions and increase safety for laborers is pushing many manufacturers to consider the imple- mentation of autonomous carts. Bastian Solutions has created its own autonomous cart called the SmartCart Automatic Guided Cart. This design uses magnetic tape on the ground to guide itself through a warehouse environment, delivering tools and products to line workers at predetermined stopping points. Modifying the guide path is as simple as pulling up the magnetic tape and reapplying it else- where (Bastian Solutions, 2017). This system may need to be modified in the airport environ- ment, since the magnetic tape may interfere with aircraft navigation equipment or avionics. Another quickly evolving application is the autonomous forklift. The industrial vehicle company, Seegrid, has produced a fully autonomous forklift that is capable of maneuvering in warehouses, lifting heavy loads, transporting them where they need to be, and unloading them (Dormehl, 2017). The capability to lift heavy materials is very useful in the manufacturing environment, and could also have airside applications for lifting baggage and cargo. Moving from the manufacturing industry to healthcare, autonomous carts or robots are increas- ingly popular in hospitals around America. Hospital robots are designed to deliver food, clean linens, and even medication to patients, and take away medical waste, soiled linens, and trash (Simon, 2015). The robot is also able to call and use elevators, maneuver around obstacles, and utter short phrases about intended actions. The hospital robot “Tug” at San Francisco’s Mission Bay Wing at the University of California is able to move up to 1,000 pounds and has security features that make medicine delivery possible. Security measures to protect unauthorized access include the use of a required pin code and fingerprint scan before a drawer will open; this ensures that only nurses with prior authorization can access the secured contents. A second security feature is that the robot can only be opened when it has reached the programmed location, which ensures that no one can access the contents en route or in an elevator (Simon, 2015). The goal of the hospital robots is not to eliminate nurses, but to allow nurses to focus on medical tasks rather than tasks such as the transport of linens, food, or trash. Patients need comfort and care from human nurses, and hospital robots allow nurses to spend more time on specialized tasks and less time delivering laundry. Automated Processes Autonomous machines have been used in car manufacturing since 1961, when General Motors used them for tasks such as spot welding (Robotics Online, 2017). Today, automated components in car manufacturing have become a necessity to keep up with strict production deadlines that would be impossible without automation. Assembly lines have been used in manufacturing for decades, and the newer addition of automated arms help increase produc- tion rates even more (Deaton, 2009).

Applications and Lessons Learned 15 Automation in manufacturing results in a safer and more efficient process than reliance on human labor. This is especially true for tasks such as welding and lifting large, heavy, or awkward parts. Elon Musk, CEO of Tesla, is vocal about his goal of complete automation in the company’s Fremont, California, factory. Musk wants to remove humans from manu- facturing altogether, replacing them with faster and more efficient automated machines. To support this goal, Tesla purchased Grohmann Engineering, a German firm that specializes in automated processes (DeBord, 2017). Beyond automotive manufacturing, the Danish company Universal Robots has created what is called collaborative robotic arms, or “cobots.” The highly fluid motion of the cobots lends itself to hundreds of applications in manufacturing. Universal Robots sells a variety of different tips, called “end-effectors” for different tasks, such as welding, vacuuming, polishing, assembly, and object analysis. The newest generation of cobots takes one hour to unpack, mount, and program with the first task (Digital News Asia, 2018). Companies may purchase the generic arm and the end piece that is suitable for the specific task they wish to automate. More than 24,000 of these collaborative robotic arms are in use around the world (Universal Robots, n.d.). The ease of use and flexibility of applications are making these robotic arms extremely popular in the manufacturing industry. Automated Perimeter Security The Indianapolis Motor Speedway security force implemented a 400-pound Sharp INTELLOS Automated Unmanned Ground Vehicle called ROSS-E in May 2018. With the capability to cruise at up to 3 miles per hour, ROSS-E uses a variety of sensors and cameras with infrared capability to scan the environment around the perimeter fence to locate trespassers and security threats, as well as to aid in obstacle avoidance. The machine features two-way communication so someone in the security office can speak through the vehicle to anyone they need to address (Guskey, 2018). The ROSS-E robot is the second robot purchased by The Indianapolis Motor Speedway for security surveillance. During races, it serves as a stand-still visibility piece, using its multiple cameras to scan everything going on. After hours, it roams the grounds regardless of weather conditions (Perez, 2018). The use of security robots has enhanced security and reduced the number of human workers that need to patrol the perimeter of the raceway. Applications in Aviation Airside activities must provide safe and efficient operations to support mobility for domestic and international passengers and goods. The strategic implementation of AGVT for airside activities may provide an important tool to allow airports to effectively meet their mission to provide the safest, most efficient aerospace system in the world (FAA, 2017a) and realize their vision for sustainable growth in global civil aviation in the future (ICAO, n.d.). Assuring safe and efficient operations while supporting growth requires innovation, forward thinking, and a willingness to try solutions that have not been attempted in the past. Recognizing the potential for AGVT, autonomous vehicles were one of the three key ideas for Simplifying the Business (StB) (IATA, 2017). The purpose of the StB report is to highlight innovation and key projects that aim to improve the passenger experience. Led by IATA, a team of key stakeholders in the aviation industry meet periodically throughout the year to brainstorm and develop innovative ideas. The fact that autonomous vehicles were highlighted in the StB report of 2019 recognizes the potential for applications of AGVT in airside operations in the near future.

16 Advanced Ground Vehicle Technologies for Airside Operations IATA identified automation as a solution for the increased demand for air travel. According to the StB report, AGVT offer, “a near-term option to improve airfield efficiency and effectiveness without significant capital outlay” (2017). In addition to supporting efficiency and effectiveness, AVs may also increase safety and reduce aircraft damage. Many of the vehicles that operate on an airfield can cause moderate to severe damage to aircraft, as seen in Figure 3. There are many potential applications for AVs and AGVT in airside operations as described below. Automated Snowplows In March 2018, self-driving snowplows were deployed in a demonstration project at Fagernes Airport in Leirin, Norway, 125 miles north of Oslo. The demonstration featured three snowplows working in formation to clear runways and taxiways, fully autonomously, without drivers in any of the vehicles. The snowplows can maintain the runways and taxiways regardless of the weather, and in the demonstration, they cleared 88 acres per hour. The technology was devel- oped by Yeti Snow Technology, which is owned by Semcon and Øveraasen for Norwegian airport operator Avinor (Airport Technology, 2018). Semcon has expertise in real-time systems and autonomous technology and designed a control system that sets up digital patterns for auto- nomous snow clearance at airports. The system can download the patterns and monitor a number of vehicles that navigate using RTK GPS for accurate position measurement and communicate using 4G modems (Semcon, 2018), as shown in Figure 4. Øveraasen is an airport vehicle developer for the 65-feet long and 18-feet wide autonomous snowplows used in the demonstration (Esterdahl, 2018). As of September of 2018, Project Yeti was expected to perform live airport operations beginning in January 2019 at Oslo Airport (Maronese, 2018). Daimler is another company working on automated snowplows. Building on their work on the Highway Pilot Connect Systems, they developed an autonomous snowplow to follow GPS-mapped routes on a runway. The snowplows can be platooned through the mapping done by the user. Operators in the lead vehicle can override control and make changes through their truck’s control panel (Daimler, 2017). At the Winnipeg Richardson International Airport, automated snow removal trails are also underway. With the assistance of Northstar Robotics and Airport Technologies, the airport will soon utilize autonomous snowplow equipment for their runways. Trials for the technology will not be performed around active aircraft and the group will follow normal Transport Canada safety protocol (CBC News, 2018). Figure 3. Damage caused by equipment. Based on data published by IATA, 2017.

Applications and Lessons Learned 17 Automated Airport Shuttles In summer 2018, Gatwick Airport in southeast England is introducing new electric-powered autonomous vehicles to shuttle airport employees. This trial is the first of its kind for an airport and aims to pave the way for future autonomous vehicles in airside operations. Gatwick’s 300 air- side vehicles are stationary 90% of the time, and the airport hopes that eventually airfield trans- port needs can be met with a much smaller fleet of autonomous vehicles. The data collected in the trial will be used in communications with the Department of Transport, Civil Aviation Authority, and others leading the future of transportation (Gatwick Airport Press Office, 2018). Figure 4. Automated snowplow demonstration at Fagernes Airport, Lerin, in Norway, March 2018, (a) conceptual diagram for control of snowplows and (b) automated snowplow platoon. Photo: Esterdahl, 2018. (b) (a)

18 Advanced Ground Vehicle Technologies for Airside Operations Unlike the autonomous airport pods that have been moving passengers between the terminal and parking at Heathrow Airport since 2011, the vehicles with AGVT at Gatwick will not be confined to a closed course (Gatwick Airport Press Office, 2018). Trials of the autonomous shuttle in an open environment at Gatwick will lay the foundation for other applications within airside operations (Gatwick Airport Press Office, 2018). Gatwick’s deployment of an airside autonomous vehicle to build knowledge makes sense as there are many possible applications for autonomous vehicles in the airside environment. IATA advocates more than 40 use cases for autonomous vehicles, including pushback tugs, passenger load bridges, and baggage vehicles. Of the 40 use cases identified by IATA, 25 are shown in Table 2. Nineteen of these uses are for airside activities and six are for terminal or landside activities. Additional information about potential uses identified by IATA is shown in Appendix B. There are a number of other airports that have or will be implementing AGVT shuttles, but most will be used landside. Brussels airport will have automated shuttles on public roads leading to the airport as well as landside on the airport and is expected to begin in 2018 (Brussels Airport, 2018). Navya, based in Lyon, France, has shuttles at a number of places including Charles de Gaulle Airport (CDG), where they provide services from the parking lot to the airport terminal (Phelan, 2018). In early 2018, IAG Cargo began official trials of AGVT at Heathrow Airport. Rather than deploy automated shuttles, some companies have chosen to incorporate advanced technologies that lay a foundation for future automation. Transdey has partnered with safety start-up, Driversiti, to deploy safety analytics software in SuperShuttles, which currently serve airport shuttle fleets in more than 40 cities in the United States and seven cities internationally (Bhuiyan, 2016). This technology will build on sensors in smartphones and tablets, and will be able to detect when a car several cars ahead stops abruptly (or gets into an accident) and provide an alert to other cars nearby that are equipped with the software. Airside Applications Landside and Terminal Applications Airside Ground Vehicles and Equipment Baggage/cargo carts Dollies and loaders Aircraft tugs Jet bridges Aircraft marshalling cars Employee/passenger buses/shuttles Baggage Airside Robots AKE and AKH robot loaders at laterals in bag hall Robot loading AKE and AKH into aircraft Mobile Robot security screen pod—bag/passenger Aircraft inspections Perimeter monitoring Lawn mowing and deicing/snow clearance Airside Deliveries Airside passenger and employee buses/shuttles Catering trucks Cleaning crew vehicles Remote PRM airplane loaders Maintenance vehicles Small parts/items/rush bag Landside Transport PRM wheelchairs PRM carts in the terminals Rental car and parking lot shuttles Employee and passenger buses/shuttles Baggage Landside or Terminal Operations Mobile security robots and kiosks Note: AKE and AKH are unit load devices certified for aircraft use, PRM is passengers with reduced mobility. Table 2. Use cases for ground-based automated vehicles as identified by IATA (2017).

Applications and Lessons Learned 19 Use of smartphones will allow the technology in existing vehicles, rather than waiting for automakers to integrate the technology into new cars. This capability to work with existing vehicles may be valuable, given the long production cycles for vehicles. Although other startups leverage the GPS and accelerometers in smartphones, most use the phone as a passive device to gather data, which is later uploaded to the cloud and interpreted. This provides limited value for safety alerts in real time although it can warn employers of erratic driving behavior. Driversiti’s system, on the other hand, utilizes the smartphone or tablet as a portable beacon and is designed to work with or without an internet connection. SuperShuttle will be the first commercial fleet to be equipped with Driversiti’s technology (Bhuiyan, 2016). Perimeter Security Building on the kind of success demonstrated by automated security at the Indianapolis Motor Speedway, Edmonton Airport in Canada (YEG) announced plans to use an autonomous all- terrain vehicle (ATV) for perimeter security in July 2018. The vehicle was developed by the Alberta Centre for Advanced Microprocessor and Nanotechnology Products and can be remotely controlled by people or drive autonomously (Sarkonak, 2018). The vehicle will use machine learning to support automated operation. In addition to security functions, the airport system has been trained to detect animals such as coyotes and deer, and when an animal is detected, security is alerted and a report documenting the species is made. The vehicle is considered to be in the development phase. Automated Construction Activities Automated construction equipment has been successfully used for numerous airfield pavements. One of the first applications was in 2010 at the John F. Kennedy International Airport (JFK) for a $204 million reconstruction of its 14,572-foot runway (For Construction Pros, 2010). This project widened one of the nation’s longest commercial runways from 150 feet to 200 feet and included milling, grading, and asphalt shoulder paving supported by automated machine controls, which reduced the milling schedule by weeks, compared to the use of tradi- tional staking and manual controls, and allowed a year’s worth of production in 100 days (For Construction Pros, 2010). Robotic paving equipment provides automated control in 3D for precise and rapid milling and paving. Equipment for a runway paving project, shown in Figure 5, includes the paving machine with robotic controls and the on-site survey equipment needed to support robotic milling and paving. Although it requires an initial capital investment, it dramatically reduces the time required for project completion. In many cases, the same advanced technology can be used to support both milling and paving activities. For example, at Keflavik Airport in Iceland, the Leica PaveSmart 3D solution was used in conjunction with the Wirtgen Milling Machine W200i for milling off the old surface and the Vögele Super 2100-3 Navitron Paver for placement of new material. Use of the same technology and equipment to support both milling and paving leverages the capital investment in the equipment, and also provides advantages in terms of personnel training (95% of the steps are identical for both milling and paving) and contracting, since a single company can complete both tasks (Leica Geosystems, n.d.). Automated 3D machine control also supports the very smooth pavement requirements needed for runways, which is 3 mm in some cases. The required tolerance for smoothness can be very hard to achieve without 3D machine control. Robotic milling and paving also eliminates the need to manually stake out control points and manually measure for grade control. This results in a faster process with fewer stops, which can reduce the duration of runway closure.

20 Advanced Ground Vehicle Technologies for Airside Operations Automated Jet Bridge Automated jet bridges may reduce aircraft damage, and a completely automated system may allow passengers to deplane from the aircraft even when ramp workers are not allowed on the ramp due to weather conditions such as lighting. Japanese company ShinMaywa makes specialty vehicles and aircraft parts and has 60% of the market for automated jet bridges in Southeast Asian airports. Their automatic jet bridge has the capability to automatically adjust the height and length of the jet bridge to maneuver within approximately 20 inches (10 cm) of an aircraft’s door (Kuroda 2018). Using information from a touch screen that indicates the type of aircraft at the end of the jet bridge, the jet bridge automatically adjusts to meet the aircraft in a process that takes only 30 seconds and is assisted by a variety of infrared sensors and high-quality cameras (Onishi, 2016). Over 190 ShinMaywa boarding bridges are in operation at Changi Airport in Singapore. In the future, the company plans to introduce more autonomous features to their jet bridges, including detection and troubleshooting of mechanical issues and incorrectly parked aircraft (ShinMaywa Industries, n.d.). Similar to ShinMaywa, Singapore Technologies Engineering has a jet bridge in development that will be able to dock to aircraft in poor weather conditions, which will reduce delays. The group is planning on implementing the devices at Changi Airport where the airport is focusing on robotics and automated systems which it hopes to have implemented within 5 to 6 years (Lim, 2017). Changi Airport’s Terminal 4 has incorporated a variety of advanced technologies inside the terminal as well as airside. Opened in October 2017, the one-billion- dollar Terminal 4 utilizes automated ticket counters and check-in areas which utilize kiosks rather than employees, even for checked luggage, which passengers place on a conveyer belt themselves; it is rare to see airline personnel in the terminal (Saiidi, 2017). As the airport begins to build their biggest terminal, Terminal 5, they are leveraging the technology tested in Terminal 4. The ultimate plan is to have the advancements working so they can be used in totality in Terminal 5 (Park, 2018). Aircraft Tug and Tow There are a number of efforts underway to use advanced technology to support aircraft tug and tow activities. Two of these include the Mototok Robotic Tug and Taxibot. Another Figure 5. Automated robotic milling and paving for airside construction: (a) robotic paving equipment provides 3D automated control for precise and rapid runway paving and (b) on-side survey equipment provides information needed for robotic milling and paving machines. (b) (a)

Applications and Lessons Learned 21 technology that will affect aircraft taxis is the use of a self-driving nose wheel with an electric motor powered by the plane’s auxiliary power unit. Major aircraft manufacturers such as Boeing and Airbus are both working on this technology, which would save time and reduce fuel consumption and emissions. WheelTug is one of the companies working on this technology in partnership with aircraft manufacturers and airlines. Safran is another company working with Airbus to develop this technology (AviationPros, 2017a). Mototok Robotic Tug (Mototok Spacer 8600). British Airways uses five Mototok robotic tugs for operations at 25 of their gates at Heathrow’s Terminal 5 (LeFebvre, 2017). These towbarless electric tugs move planes with more precision than traditional driven tug vehicles, can handle up to 25 pushbacks with one battery charge, can recharge in 3 hours, and have minimal operating and maintenance costs. The robotic tugs are not yet fully autonomous, but are remotely operated by joystick, and the operator must maintain visual line-of-sight to ensure obstacle avoidance. The operator can be seen walking with the joystick in his hands in Figure 6. British Airways noted that the use of this robotic tug helps ensure on-time perfor- mance, and is part of a targeted effort to modernize operations and use technology to enhance customer experience (LeFebvre, 2017). TaxiBot. The efficient movement of airplanes from the gate to the runway is a priority at many airports, especially where high demand results in delays. According to Airbus, this operation alone produces 18 million tons of CO2 per year at a cost of $7 to $8 billion a year (Cummins, 2009). Use of aircraft jet engines on this movement also increases the risk of foreign object debris (FOD) damage, which costs $350 million a year, due to aircraft maintenance and delays (Cummins, 2009). Solutions that allow engine-off taxiing to the runway would provide benefits such as reduced fuel use, reduced emissions and environmental impact, reduced noise and more efficient operations. In 2006, Virgin Atlantic attempted to address this and created airport “starting grids” that would allow planes to be towed close to the runway. This program was discontinued due to design flaws that caused maintenance problems with the landing gear due to continual jerks inflicted by the towing vehicles (Cummins, 2009). TaxiBot was another solution proposed in 2009. TaxiBot allows the aircraft pilot to remotely control the tug to transport the aircraft to the departure runway from the gate. TaxiBot was deployed in a trial with Lufthansa at Frankfurt Airport and received certification from European Aviation Safety Agency (EASA). TaxiBot can travel at speeds up to 23 knots and saved 50 to 100 kilograms (110 to 220 pounds) of fuel for each taxi-out operation of a Boeing 737 (GreenAir Figure 6. Technologies for aircraft tug and tow: (a) ramp agent using Monotok robotic tug and (b) TaxiBot at Frankfurt Airport in 2015. Photo: Monotok, n.d.; TaxiBot, 2018. (b) (a)

22 Advanced Ground Vehicle Technologies for Airside Operations Online, 2015). Although the trial was not extended, there is interest in the technology with deployment plans for airports in India and simulation studies by airports such as Charles de Gaulle to determine the potential benefits and compatibility with infrastructure. The Charles de Gaulle Airport study (with Air France) found that TaxiBot is compatible and would have no adverse effects on aircraft flows. In 2018, Delhi International Airport announced that they would be implementing 17 TaxiBots at the Chhatrapati Shivaji Mumbai Airport (BOM) and Indira Gandhi New Delhi Airport (DEL) over the next 4 years (Gandhiok, 2018). In an effort to move to markets around the world, TaxiBot aims for approval from the Civil Aviation Administration of China (CAAC) and has received EASA and FAA certification for both the Boeing 737 family and the A320 aircraft family; this encompasses over 70% of the commercial airline fleet. TaxiBot has designed their tug to include a safety driver in the tug cabin (GreenAir Online, 2015), and although current tugs are not entirely automated, they provide an important stepping stone towards the future of increased automation for airside vehicles. TaxiBot is also developing a system that is fully electric. A research study conducted in 2015 suggested that a fully autonomous aircraft tug would be a good solution to safety issues and inefficiencies with aircraft taxiing (Morris et al., 2015). The proposed tug is instrumented with LiDAR and two electro-optical/infrared cameras, one front-facing and one rear-facing. The front camera is used for recognizing runways, taxiways, and gate markings for navigation. The rear camera is used for docking/undocking and monitoring the state of the aircraft during towing operations. Despite the well-defined area of an airport, video test data collected at the South Jersey Regional Airport highlighted many infrastructure problems that could affect the automated tug. These include eroded and weathered paint, pave- ment stains and discolorations, cracks, and false edges. The proposed tug may have a difficult time recognizing and maneuvering around these imperfections that are commonly found at airports (Morris et al., 2015). Research at Atlantic City International Airport using the Velodyne HDL-64E LiDAR sensor found that the use of reflective markings on vehicles, cones, and vests worn by airport personnel intensified the LiDAR readings (as seen in Figure 7) and thus would make it easier for an autono- mous tug to operate. The figure shows aircraft and ramp equipment from the tug point-of-view. The increased use of reflective markings is often already being implemented at airports to improve safety, and this may support the future use of LiDAR equipment to identify obstacles and operate in the airport environment. These reflective surfaces may also decrease the need for improved infrastructure, as the LiDAR can detect enough details to navigate itself through the environment without it being in excellent physical condition. Figure 7. Lidar image at Atlantic City International Airport. Photo: Morris et al., 2015.

Applications and Lessons Learned 23 Any automated tug system must also be able to respond to unforeseen events such as mechanical trouble, unexpected obstacles, and communications failures. Research is underway to address these issues to assure that an automated tug system can handle such contingencies (Morris et al., 2015). Investigation of a system for autonomous towing found that four major requirements must be met (Morris et al., 2015): 1. The autonomous tugs are safe, avoiding people and obstacles. 2. There is minimal disruption to normal operations; humans do not have to change their behavior much. 3. No major airport infrastructure has to be redesigned; the autonomous tug system can work on existing taxiways and ramp areas. 4. Use of the autonomous tugs only improves operations and helps humans be better at their jobs. Additionally, there were three kinds of challenges to consider when attempting to add autonomy to airport surface operations (Morris et al., 2015): 1. Technical challenges: the autonomous system must be intrinsically safe, able to adapt to changing environments, and reliable. 2. Economic challenges: the autonomous system must be affordable, have no external hidden costs to the customer, and provide a strong business model. 3. Social Challenges: the autonomous system must offset any potential job loss with a positive impact on the labor force, must operate in a safe and familiar way to humans. A system for autonomous towing remains under investigation. Charlotte Douglas Inter- national Airport has been identified as one potential airport for deployment of autonomous tugs due to the use of the Airspace Technology Demonstration 2 (ATD-2). Other airports such as Dallas/Fort Worth International Airport will soon utilize ATD-2 and may also be appro- priate candidates. San Francisco International Airport has been identified as another potential airport for deployment of autonomous tugs, due to the congestion, the numerous carriers, and the widespread acceptance of advanced technology in the San Francisco area. In 2018, NASA, Lufthansa, RS&H, and San Francisco International Airport submitted an autonomous taxiing concept to the IATA innovation contest (DiPrima and Fong, 2018). Aircraft Parking Guidance System Airports have been using technology to aid aircraft parking since the 1970s (Phippen, 2014). It began with the Standard Visual Docking Guidance System (VDGS), which guides pilots to the ideal parking location at their gate using both human and mechanical direction. The tech- nology has progressed, and manufacturers now provide more autonomous solutions for aircraft parking. Referred to as Advanced Visual Docking Guidance System (A-VDGS), these systems use lasers to identify the type of aircraft approaching the gate and guide them with visual prompts on the multi-color LED display into the best parking orientation (Honeywell, 2013). This is a completely autonomous solution. Ramp personnel do not need to be present for the pilots to self-park, which is especially helpful in the event of a thunderstorm or in other conditions where personnel are advised to stay indoors. The Safedock A-VDGS parking guide that pilots see is shown in Figure 8. This technology and its benefits have been proven for over a decade. As early as January of 2008, 92 of the 157 gates at Dallas/Fort Worth International Airport used the advanced Safedock system for automated aircraft guidance into gates (Richards, 2008). The initial investment to install the self-docking system cost $6 million, but everyday use in normal weather condi- tions saves airlines approximately $3.5 million a year in fuel and increased capacity. The system

24 Advanced Ground Vehicle Technologies for Airside Operations also allows operations during lightning conditions, which further increases capacity and yields an additional $11 million in savings for the airlines. Altogether, the implementation of these automated docking systems saves airlines a total of $14.5 million annually (equivalent to $17 million in 2018). Beyond monetary value, the system is reducing aircraft emissions, delays, and providing a better customer experience (Richards, 2008). Automation in Cargo and Baggage Technology to support automation of cargo and baggage movement has been initiated by a number of airports. Lufthansa Cargo has collaborated with industrial machine vision company VITRONIC to implement an automated solution for analyzing and documenting their air freight. Ten automated volume scanners have been installed at six of the airline’s hubs around the world, which quickly scan the outbound cargo crates and determine their exact dimensions and weight. The data and photos are then uploaded to Lufthansa’s IT system, where it can be referenced and used for documentation or billing (Mänz, 2016). This technology has the potential to make way for many new advancements for Lufthansa, such as algorithmic organization of the cargo within the aircraft and strengthening their global network. The CargoPod vehicle, produced by automated car company Oxbotica, ran autonomously along a cargo route around the airside perimeter for 3½ weeks. The goal of the trial was to assess the applications and future potential for the automated car. They collected data from 200 km of vehicle travel that they will use in their assessment (Air Cargo News, 2018). Auto- mated cargo movement is also being tested at Changi Airport (SIN). In addition to moving cargo, the trial investigates the transport of cargo documents delivered using a driverless vehicle at Changi Airport (Park, 2018). In terms of automation to support baggage movement, Rotterdam The Hague Airport is working with logistics company Vanderlande to deploy a new baggage handling system, with trials in 2018 (Airport World, 2018b). The system will not use conveyer belts per a traditional Figure 8. Safedock A-VDGS. Photo: ADB SAFEGATE, 2018.

Applications and Lessons Learned 25 sorting system but rather will deliver individual bags directly to their destination using auto- mated carts (International Airport Review, 2018). A cart, a small robotic vessel, will retrieve the bag and move to where it needs to go. A 50% increase in energy savings is expected, as compared to a traditional baggage system (International Airport Review, 2018). Another expected advantage is that problems would result in the delay of a single bag rather than a shutdown of the entire system. The implementation began in November 2018. Dallas/Fort Worth International Airport announced a partnership with Vanderlande in the Summer of 2019 using the carts shown in Figure 9 (Dallas/Fort Worth International Airport, 2019). The airport will be using Vanderlande’s autonomous carts for baggage in Terminal D; the system can handle 450 bags per hour and is being tested for connecting international passengers. Passengers will use self-service bag drops and identify the airline for their connecting flights. The FLEET system will transport each piece of luggage to the appropriate bag belt via autonomous vehicles (Dallas/Fort Worth International Airport, 2019). Although the carts shown are used inside the airport, once the technology is demonstrated in this environment, it may be possible to expand the concept airside. Companies in the AV sector expressed interest in automated baggage applications due to the potential opportunity to gain knowledge that would be helpful for delivery in the roadway environment, which is a substantial market. Advanced Technologies for Air Traffic Control and Airport Ground Surveillance Advanced technologies to support airport ATC and airport ground surveillance include the Airport Surface Detection System—Model X (ASDE-X), Airport Surface Surveillance Capacity (ASSC), and Airspace Technology Demonstration 2 (ATD-2). ASDE-X and ASSC. ASDE-X and ASSC are real-time airport ground surveillance systems that use radar, multilateration, and satellite technologies to track the surface movements of aircraft and ground vehicles on the airfield. This information is translated to a visual interface that overlays an airfield map. The primary goal of this technology is to reduce critical Category A and B runway incursions by alerting ATC to any hazardous movements (FAA, 2014a). Figure 9. Automated baggage carts transport individual bags at Dallas/Fort Worth International Airport. Photo: Dallas/Fort Worth International Airport, 2019.

26 Advanced Ground Vehicle Technologies for Airside Operations Airports with ASDE-X and ASSC are able to track both transponder and non-transponder equipped vehicles in the movement area, and can utilize ground vehicle ADS-B transponder data, as well as data from other sensors or surveillance feeds. If ground vehicles use ADS-B transponders, there will be additional information regarding specific vehicle characteristics (e.g., identification, speed, and similar information to ADS-B equipped aircraft). These systems are already being used at 36 of the busiest airports in the United States. ASDE-X and ASSC pro- vide controllers with color displays to provide accurate and reliable surveillance in all weather conditions. In addition to showing current conditions, displays can provide data recording and playback, as well as system status monitoring and decision support tools (Maum, n.d.). Utilization of ground vehicle ADS-B transponders [aka ADS-B out squitter, ADS-B squitter units, ADS-B squitters, Vehicle Movement Area Transmitters (V-MAT or VMAT)] enhances situational awareness for both aircraft and vehicles, and ADS-B squitters can be funded though the AIP (or using PFC) for eligible airports. Real-time maps that reflect the current location of nearby aircraft and other ground vehicles can be displayed in a moving map in the ground vehicle. FAA has approved both V-MAT and external mount V-MAT (FAA, 2016a). Ground vehicle ADS-B squitters can be transferred to different vehicles, as needed. For example, the same ADS-B squitter can be used on a snowplow in the winter, and a construction escort vehicle or mower in the summer (C. Zanardi, personal communication, September 17, 2019). Each ground vehicle has its own call sign and ICAO code and each VMAT being used needs to be reprogrammed with the proper vehicle ID and ICAO code when seasonal transfer of the VMAT is performed (C. Zanardi, personal communication, October 2, 2019). The external mounts manufactured by FreeFlight Systems are no longer being produced; the only FAA approved VMAT currently available are the internal mounts provided by L3Harris Technologies (L3Harris) (C. Zanardi, personal communication, October 2, 2019). Airports that do not have ASDE-X or ASSC can use available L3Harris FAA ADS-B surveil- lance (if available) or install Symphony® ADS-B Xtend to provide surface tracking of aircraft and ground vehicles (Harris, 2016; C. Zanardi, personal communication, September 17, 2019). Symphony ADS-B Xtend can also be used to mitigate gaps in surface coverage at airports with ASDE-X or ASSC, and can be deployed on a temporary basis, if needed (Harris, 2016). Limits regarding the use of ADS-B for airside vehicles are discussed in the fleet management section. ATD-2. NASA has developed ATD-2 to enhance efficiency and communication. According to NASA, “Together with the FAA and industry, NASA’s ATD-2 integrates arrival, departure, and surface (IADS) concepts and technologies to demonstrate the benefits of an IADS traffic management system for Metroplex environments” (NASA, 2016). The first airport to utilize ATD-2 was Charlotte Douglas International Airport, which was chosen due to its size and the number of operations (Gipson, 2017). Three separate FAA technol- ogies are combined in this initiative: traffic flow management system, time-based flow manage- ment, and terminal flight data management. The ATD-2 systems analyze historical and real-time data to support more efficient scheduling and sequencing of aircraft. There are several benefits to using the system, but NASA is focusing their efforts on reducing delay for the movement between gate pushback and arrival at the departure runway. ATD-2 with terminal departure scheduling reflecting operations at multiple locations will be implemented to support opera- tions at Dallas/Fort Worth International Airport and Dallas Love Field Airport (NASA, 2016). Advanced Technologies for Ground Vehicles There are a variety of technologies to support airport ground vehicles. Some of these tech- nologies, such as collision warning systems, are commonly used in the roadway sectors, although

Applications and Lessons Learned 27 the operating parameters (e.g., recognition of the runway threshold and deployment of asso- ciated warnings) are tailored to the airside environment. Applications discussed below include fleet management, collision warning systems, and advanced sensors for airside vehicles such as airport rescue and firefighting (ARFF) equipment and snowplows. Fleet Management. Fleet management uses GPS and provides information regarding the location, movement, and current status of airport equipment. Fleet management software can provide information that improves the efficiency of equipment and allow managers to make decisions that lead to a greater operational efficiency. In the aviation sector, there are a variety of fleet management options. Fleet management systems are well developed and have been deployed successfully in a variety of sectors. For ramp equipment, one option is Fleet Online Airside, a program developed by AdaptaliftGSE that uses sensors to identify the current position and track each piece of ground service equip- ment (GSE) equipped with a transponder. The software provides driver verification, vehicle control and monitoring, geofencing, safety overrides, and supports mobile communications (AdaptaliftGSE, n.d.). The equipment is ISO 9001 certified as a registered quality management system. Other systems include Avro’s GSE Tracker, which provides users with a fleet overview, scheduling, control and override technologies, email alerts, and training reminders or sugges- tions (AVRO GSE, n.d.), and Inform’s GroundStar, which can track airport equipment and vehicles as well as provide data to identify safety issues and concerns (INFORM GmbH, n.d.). Fleet management utilizing ADS-B transponders is one option for airport operations vehicles and this option integrates operational data for aircraft and ground vehicles, and increases situational awareness for both pilots and ground vehicle operators. The FAA currently limits the use of ADS-B transponders for ground vehicles. The restrictions are as follows (FAA, AC 150/5220-26, 2011): • Only vehicles that operate in the movement area (e.g., the runway and/or taxiway) can use ADS-B squitters. • Only airport vehicles, ARFF, FAA vehicles, and other approved vehicles can use ADS-B squitters. Other approved vehicles may include aircraft tugs owned by airlines at some airports (C. Zanardi, personal communication, October 2, 2019). • Only 200 ground vehicle ADS-B squitter units per airport are allowed by FAA. Approval is underway regarding the use of ADS-B ground vehicle transponders in the non- movement area, and in the future it is likely that ground vehicle transponders can be used in the non-movement area as well as the movement area (C. Zanardi, personal communication, September 17, 2019). The current limits prohibiting use in the non-movement area make it impractical for ramp vehicles (e.g., baggage carts) to use a fleet management system based on ADS-B transponders; however, ramp vehicles could use a fleet management system that receives ADS-B information from an aircraft and integrates it with other technologies for the ground vehicle fleet. Collision Warning and Runway Incursion Warning Systems for Ground Vehicles. Collision warning systems are becoming a standard feature for vehicles; however, the benefits of these systems have not been documented in the airside environment. A literature review of tech- nological solutions to prevent surface accidents and runway incursions was conducted by the FAA Airport Technology Research and Development Branch a number of years ago (Doig et al., 2012). Although technology has progressed since then, this review included an assessment of three major technologies: • Incursion collision avoidance system (ICAS), • Runway incursion monitoring detection alerting system (RIMDAS), and • Asset tracking and incursion management system (ATIMS).

28 Advanced Ground Vehicle Technologies for Airside Operations The results found that the most readily implemented technology was the ATIMS, which was the only system that did not require the installation of additional equipment on the airfield. The ATIMS program includes a map that shows all vehicle locations, speed, direction of travel, and location history, as seen in Figure 10. ATIMS is capable of showing both real-time and historical location, speed, and direction data for airfield vehicles. The program can be configured with sensitive areas, often around construction or non-movement areas, which set off a warning to the driver when they are approached (Doig et al., 2012). This technology does not warn drivers of a potential collision with other ground vehicles. However, current vehicle technologies have been widely implemented in newer cars and they do include collision avoidance systems. The Insurance Institute for Highway Safety has added collision avoidance system testing to its list of safety evaluations due to the strong potential for an increase in safety (Linkov, 2015). Draft revisions to AC 150/5210-25A, Performance Specification for Airport Vehicle Runway Incursion Warning Systems (RIWS), were published in December 2018, for review and feed- back. Changes to the advisory circular include the following: • Alarms have been redefined as a caution alarm and a warning alarm with audibility levels and duration defined per the Human Factor Design Standard (Ahlstrom and Longo, 2003). • Visual alarms have been redefined as per MIL-STD-1427G, Department of Defense Design Criteria Standard, Human Engineering (2012). • A fixed proximity zone has been defined, replacing the need for speed and proximity warning calculations. • RIWS performance specifications have been updated for compliance with RTCA/DO-280, Minimum Operational Performance Standards for Airborne Supplemental Navigational Equipment Using Global Positioning System (GPS) (1991). • Performance requirements have been established for RIWS mobile apps. Figure 10. Asset Tracking and Incursion Management System. Photo: Team Eagle, 2017a.

Applications and Lessons Learned 29 Information from the newly referenced documents that are incorporated into the draft advisory circular may be useful as guidelines and eventually standards for other AGVT appli- cations airside. Cameras and Infrared for ARFF Vehicles and Snowplows. AGVT includes traditional cameras and infrared cameras to provide drivers with enhanced situational awareness of their surroundings. The value of these technologies is recognized by its inclusion in AC 150/5220-10E, Guide Specification for Aircraft Rescue and Fire Fighting (ARFF) Vehicles (2011), which requires a forward-facing infrared camera that provides better visibility, allows firefighters to find a route in zero visibility conditions due to smoke, darkness, or fog, and locate and avoid accident survivors. Newer products not only improve performance, but also enhance ergonomics and keep firefighters further away from the fire (Oshkosh Airport Products, 2015). Lessons Learned Aviation may leverage lessons learned from other sectors, when it comes to future deployment of AGVT in the airside environment. Lessons learned from UAS at airports, AVs in the roadway sector, and previous technology deployments at airports should all be considered. Unmanned Aerial Systems at Airports An examination of the application of UAS at airports illustrates some of the possible issues and opportunities for AVs at airports. The applications of UAS and AVs have similarities at airports. Both utilize automation in a highly regulated environment controlled by FAA and both require communication and coordination from multiple parties for safe operation. Both of these technologies can also potentially provide unique solutions to improve operations and capabilities at airports. There has been a steady increase in popularity of both UAS and AVs in the last few years, and both technologies are expected to rapidly advance in the near future. There are also important differences between UAS and AVs. UAS have unique operating restrictions due to airspace considerations. They are limited in terms of the height and distance they can traverse because of potential danger to aircraft operations. In some cases, AVs could potentially conduct similar tasks without introducing issues related to airspace restrictions. These tasks potentially include runway and taxiway data collection and imagery to support airport planning and operational activities. However, in many cases UAS are more widely avail- able, less expensive, and have a higher technology readiness level. A variety of potential UAS applications have been identified for airports. UAS can support airport operations including obstruction analysis, pavement condition assessment and inspec- tion, airfield light inspections, wildlife management, security, emergency response, and construc- tion safety inspections and monitoring (Hubbard, 2017). The Savannah/Hilton Head International Airport (SAV) has collaborated with Woolpert and FAA to integrate UAS into daily operations for wildlife management and perimeter inspec- tion, which includes inspection of 12 miles of perimeter fencing that traverses swampland and is in close proximity to dangerous wildlife (Wysocky, 2018). The UAS operations program brought about required close coordination with ATC to determine suitable operating times, areas, and conditions. The UAS were required to operate within line-of-sight of the air traffic controllers free from radio frequency interference, which limited the UAV operational area. It was also necessary to create emergency response protocols in case something went wrong during operation. SAV officials were able to geofence the airport to keep the UAS in a defined area. With these operational controls, SAV demonstrated that UAS can be used in an efficient

30 Advanced Ground Vehicle Technologies for Airside Operations and safe way to conduct operational tasks, such as managing wildlife hazards and performing airfield inspections. The usefulness of the UAS program was highlighted during test flights in January 2018, when the UAS identified that a fallen tree had broken through the perimeter fence on the north end of the airport in an area that was very difficult to reach on foot or by watercraft. The UAS safely and quickly relayed the GPS coordinates of the fallen tree to airport operations personnel. Woolpert and SAV are optimistic about the future of the UAS program. Future goals include expanding the UAS operating area, obtaining FAA approval for additional activities, and refining the UAS sensors, processes, and data management systems to better suit airport needs. The ultimate goal is to use UAS for regular operations as opposed to use for a specific mission. For example, in the future the UAS could actively monitor pavement conditions and upload work orders into Cityworks. In this scenario, the data collected by the UAS would support detailed and automated record keeping, and allow airport personnel to spend their time addressing safety issues. An important lesson learned from UAS at airports is that regulatory challenges should not be considered insurmountable obstacles, as evidenced by the success of airports that have been able to successfully integrate UAS. Safety in the Roadway Sector The safety, hazards, and risks of AGVT on roadways may provide insight into the safety, hazards, and risks in airside operations. As AGVT advance and higher levels of automation become commercially available, there will be more information about the common causes of AGVT incidents and accidents, which allow support mitigation of the associated risks. Several risks have been identified from accidents within the last couple of years. These risks include automation not responding as expected, drivers not paying attention to changing road condi- tions, unexpected obstacles materializing in front of the vehicle without adequate time for the sensor recognition, and sensors unable to discern obstacles due to unusual environmental cues. It appears that sensor failure, software misinterpretation of obstacles in the environment, and operator human error may all contribute to accidents. How these risks translate to the airside environment in different applications must be studied extensively and mitigated to acceptable levels. An examination of selected crashes in the roadway sector is provided below. Tesla Crash in Florida. The May 2016 crash of a Tesla Model S in autopilot mode in Florida resulted in a driver fatality (the only occupant) when the car hit an 18-wheel tractor-trailer. The Tesla failed to brake in time when the trailer turned in front of it; the autopilot camera did not “see” the white tractor-trailer against the brightly lit sky (Tesla, 2016). As a result, the brakes were not applied before impact. This illustrates a number of potential risks, including the inadequacy of sensors in different environmental conditions and the driver’s overreliance on automation and associated inattention. The Tesla autopilot feature is actually an ADAS feature that provides speed and steering control during highway driving. Arguably the term auto- pilot may have contributed to the driver’s overestimation of the system’s capabilities. This crash highlights technical limitations, in this case unknown and limited system capabilities in this environmental and operational context, and human factors limitations, in this case the driver’s inability to respond as needed to avert an accident. NTSB attributed the accident to the truck driver’s failure to yield, inattention and overreliance on automation by the driver of the Tesla, and a lack of adequate vehicle safeguards to prevent operation on this kind of roadway (NTSB, 2017). Tesla Crash in California. The March 2018 crash of a Tesla Model X on autopilot occurred in California when the vehicle inexplicably sped up and drove directly into the median barrier,

Applications and Lessons Learned 31 killing the sole human occupant (NTSB, n.d.). A preliminary report from NTSB indicates that the driver was using autopilot with the speed set to 75 mph in a zone with a speed limit of 65 mph (NTSB, 2018). The Tesla entered the paved gore area (triangular area noted by diverging white lines) that separated the main travel lanes from an exit ramp and crashed into the crash attenuators at the end of the concrete median barrier. The impact rotated the Tesla, which then collided with two other vehicles. It appears that the autopilot system misinterpreted the pave- ment markings. This is illustrated by a video created by another Tesla driver of operation in autopilot mode on this same roadway segment where the accident happened (Lambert, 2018). The autopilot system mistakes the white lines to the left as a continuation of their current lane, and the car slowly turns left to follow the white line, heading directly for the crash attenuator in the gore area. According to NTSB, the driver’s hands were not on the steering wheel in the 6 seconds prior to the crash, and the vehicle speed increased from 62 to 70.8 mph with no pre-crash braking or evasive steering. Four seconds before the crash and at the time of the crash, the vehicle was no longer following a lead vehicle (the lack of a lead vehicle may have caused the speed increase). The accident scene is shown in Figure 11. Pedestrian Fatality in Arizona. In March 2018, a self-driving Uber Volvo XC90 operating in autonomous mode struck and killed a woman in Tempe, Arizona. The car was driving approximately 38 mph and had a safety driver present in the vehicle at the time of the accident. According to Uber, the backup driver had weeks of training and was there to take over operation of the car if needed. Footage shot from a camera in the front of the car illustrates that the pedes- trian is hard to detect, since she was wearing dark clothes and was not illuminated by lighting. The Tempe Police initially stated in a news conference, “It’s very clear it would have been difficult to avoid this collision in any kind of mode (autonomous or human-driven) based on how she came from the shadows right into the roadway” (Said, 2018). However, subsequent footage released of the backup driver inside the car moments before the accident show her looking away from the road several times (Laris and Siddiqui, 2018). It was then revealed that the backup driver had been watching TV on her cellphone for nearly an hour leading up to the accident (Korosec, 2018). According to the NTSB preliminary report, the vehicle had a self-driving system that was under development with forward-facing and side-facing cameras, radar, LiDAR, navigation sensors, and a vehicle-based computing and data storage unit (2018). The system observed the pedestrian 6 seconds before impact and classified the pedestrian as an unknown object, as a vehicle, and then as a bicycle and began braking 1.3 seconds before impact. The system is not designed to alert the driver, and the driver engaged the steering wheel less than a second before impact and began braking less than a second after impact. This crash emphasizes the risks and challenges that automated systems may have when they do not recognize a system component or are faced with an unknown or unexpected obstacle or situation. Figure 11. Accident scene of Tesla in California in 2018. Photo: NTSB, n.d.

32 Advanced Ground Vehicle Technologies for Airside Operations An important lesson learned from AV safety in the roadway sector is that AV systems do not always behave as expected, are not always able to interpret environmental cues appropriately and operator vigilance may be difficult, especially for L3. Operator overreliance on automation may become a problem as advanced technologies are deployed. Technology Deployment at Airports Airports and airlines must balance the potential benefits of being early adopters of new technologies that will improve operational efficiency with the risks of deploying technolo- gies before they are adequately developed. Most airports and airlines seem to be success- fully balancing these competing objectives through the use of demonstration projects and staged implementation of new technologies. Denver International Airport Baggage System. As construction on Denver International Airport came to a close in 1994, they were at the forefront of innovation and airport design. The airport had several impressive design aspects, but not all of them were successful in operation. DEN was on the forefront of technology with an automated baggage system that had 22 miles of twisting tracks to move the bags from check-in to the aircraft, and from the aircraft to baggage claim with as little human interaction as possible. The goal was to minimize aircraft turn time and increase overall efficiency at the airport (Harden, 2015). Unfortunately, the baggage system turned out to be a failure. The machine would launch, rip, and even destroy luggage, causing both operational issues and serious financial impacts. The failed luggage system delayed the opening of the airport by 16 months and cost the City of Denver $1.1 million per day during this time. The machine’s designers underestimated the complexity and did not create adequate backup and recovery systems to handle discrepancies and failures. Attempts have since been made to resurrect the system, but with costs ballooning from the original $200 million to some $600 million, it was abandoned entirely in 2005 (Harden, 2015). Heathrow Terminal 2. Building a new airport terminal is an enormous challenge and the integration of advanced technologies increases the complexity and risks. Heathrow Terminal 2 is an excellent example of technology success. The successful deployment of technology at Terminal 2 reflected extensive testing and a staged implementation for passengers, airlines, and the baggage system, prior to operation at full capacity (BlueSky, 2015). Terminal 2 started by serving 6,000 passengers on its first day of operation, and operated for 3 weeks at this level, which is 10% of its capacity (BlueSky, 2015). This allowed problems to be resolved with minimal impact. The terminal gradually transitioned to full service for 26 airlines, which were moved into the terminal over a 6-month period. More than 180 trials with 14,000 volunteer passengers allowed every component of the system to be tested, from parking and wayfinding to concession menus and the baggage system (BlueSky, 2015). Moreover, the system was tested beyond its limits to ensure success. For example, on average, the baggage system is expected to handle 2,400 bags an hour during peak periods; however, the system was successfully tested to handle up to 4,000 bags an hour prior to opening (BlueSky, 2015). In addition to tests in the new facilities, the airport created a model airport off-site where new technologies were tested, proven, and implemented to reduce the risk to existing operations. Extensive training assured that all employees were familiar with the system, which included new automation in a variety of applications from common use tickets to mobile technology in the terminal for passengers and mobile apps for cargo handling. One benefit of the new technology is increased opportunities for collaborations across airlines, the airport authority, and employees. The staged and successful opening of Terminal 2 contrasts dramatically with the challenges encountered upon the opening of Terminal 5 to more than 40,000 passengers on its first day, accompanied by dozens of cancelled flights, problems with staff training, and prob- lems with the baggage system (Lo, 2014). An important lesson learned from previous technology

Applications and Lessons Learned 33 deployments at airports is the value of staged technology deployment, early integration of tech- nology in the planning process, and confirmation that the technology readiness level is consis- tent with the proposed project. Other Considerations As new AGVT are utilized in the airside environment, there are a number of potential issues. These are ethical, legal, and social concerns; accident reporting; and cybersecurity. Other topics for consideration include the expected timeframe for deployment and the research underway in other transportation sectors, such as transit and roadway. Legal, Ethical, and Social Concerns There are numerous ethical, legal, and social concerns related to AV; however, many of these issues are more pressing in the roadway sector where AVs interact with the public directly. Issues such as how AGVT decision-making algorithms should address “the trolley problem” may reso- nate on public streets but may have more limited implications on the airfield (the trolley problem addresses how decisions should be made when any action taken may result in an injury or fatality). Similarly, the legal issues for AGVT in the roadway sector may be significant due to the wide range of state legislation, but the legal issues for AGVT in the airside environment probably center more on the allowable uses of AV as approved by FAA and other federal regulators such as TSA. Legal issues for AGVT in the airside environment certainly include liability. Although insurance policies differ and legal limits for public and government organizations such as airports may vary by state, it would be reasonable to have insurance riders to cover AGVT. It is ultimately the language used in the contract or policy that determines the meaning and implications of coverage. In all cases, liability may ultimately be determined by arguments in court (Prather, 2015) and the subsequent case law. Although AGVT liability coverage is not currently specifically offered in the roadway sector (Perfetto and Blancher, 2018), it presumably could be included in policy riders and may become a standard component in the future as applications become more common. Many of the ethical and social issues related to airside deployment of AGVT relate to labor force issues and the social implications of replacing human capital with automated equipment, as well as capability of the system to overcome failures. Many of these issues are theoretical at this point in the planning process, and may not need to be addressed immediately, although it is important to acknowledge that issues exist and may become increasingly important as technology progresses and applications increase. Figure 12 illustrates the conceptual overlap of ethical, social, and legal issues related to AV and CV deployments. Accident Reporting and Safety One of the best ways that accidents can be prevented is by studying accidents and near misses that have already happened. There is great value in the reporting of incidents and accidents so that all stakeholders in aviation can learn from each other. The FAA has several systems in place to encourage self-reporting and the sharing of accident information, including the Avia- tion Safety Information Analysis and Sharing (ASIAS) System and the publication of reports regarding trends and suggestions for accident mitigation (FAA, 2018a). ASIAS also partners with the Commercial Aviation Safety Team (CAST) to monitor current risks, evaluate the success of deployed mitigation techniques, and identify emerging risks. Both ASIAS and CAST focus on historic accident information to make informed decisions and recommendations to help avoid accidents in the future (Duquette, 2016). Sharing information related to the safety of AGVT will foster technology advances and will benefit all stakeholders. Having a consistent and shared record of reported AGVT accidents in the airside environment would be beneficial to further advance AGVT and to mitigate any associated risks. In the roadside

34 Advanced Ground Vehicle Technologies for Airside Operations environment, many states do not require separate reporting for vehicles with automated or connected technologies. California, however, does require reporting of any accident involving AVs as well as vehicle disengagements. A disengagement is when the car transitions from autonomous operation to control by the human driver. In most cases, this does not lead to an accident, and in some cases, a disengagement is planned as part of testing protocol. In any case, reporting information about AVs is vital for understanding how autonomous systems process information and decide to switch from an automated to a manual operational mode. A standard template is not required by California, so it can be difficult to compare data from different auto manufacturers, but at least there are records available detailing the accidents and disengagements. Table 3 shows the number of accidents and Figure 13 illustrates disengagement data reported in California. It is important to retain records of both disengagements and accidents that involve advanced technologies, as there are valuable lessons to be learned from the factors that contribute to both. If an airport deploys AGVT, it would be valuable to develop a standard template for reporting accidents, near misses, disengagements, and other pertinent information. Progress is being made toward sharing data amongst all AGVT stakeholders. In March 2018, the U.S. DOT hosted a Public Listening Summit on Automated Vehicle Policy. The department engaged stakeholders in the AV ecosystem, and found that data is a limiting factor in AV integra- tion. Data exchange is therefore vital in accelerating safe integration of AVs in the United States. The U.S. DOT then published Draft Guiding Principles on Data for Automated Vehicle Figure 12. Legal, ethical, and social implications of CVs and AVs. Based on information presented by Surakitbanharn, 2018. Year Number of Accidents Reported 2018 (through September 24) 46 2017 29 2016 15 2015 9 2014 1 Based on data reported by California DMV, 2018. Table 3. Automated vehicle accidents reported in California.

Applications and Lessons Learned 35 Integration in July of 2018. These guiding principles, shown in Appendix C, aim to assist companies, as well as government entities, in identifying their roles in data exchange. This is a step toward comprehensive guidance for companies regarding their responsibility in the exchange of AV data. The sharing of AGVT data and information related to the deployment of AGVT at a single airport would benefit all stakeholders at that airport, and sharing AGVT data and information from multiple airports would provide additional benefits. Collaborative models for information sharing have met with success. In the context of data sharing at a single airport, the Brussels airport is moving toward a collaborative and efficient operations environment with the creation of their Airport Operations Center (APOC), which includes 140 operational employees from every employer and stakeholder at the airport in an open physical location. The goal is to have open communication between all parties to increase overall efficiency and response time to emergencies. The center, which opened in 2015, has already made a great impact on the Brussels operations and has plans for further implementation in the future (Feist, 2018). Cybersecurity Cybercrime has increased exponentially in the last few years, and airports are especially vulnerable to cybersecurity attacks due to the necessary interconnectivity of their systems (Kohli, 2018). Airports are already recognizing this weakness, and are actively working to combat it. For example, Munich Airport has created a center called the Information Security Hub (ISH) with the purpose of protecting the airport and airlines from cybercrime. The center houses a team of IT specialists who work with European aviation industry experts to develop strategies for defending against cyberattacks (Airport World, 2018a). Initiatives like this will be important to assure that AGVT are safe and secure. Potential cybersecurity risks for AGVT often relate to the communications component (Petit and Shladover, 2014). Threats include GPS spoofing or jamming, in which inexpensive devices Figure 13. Sample data: (a) reported disengagement data for GM Cruise in California in 2016 and (b) sample disengagement data. Based on data reported by California DMV, 2018. Company, year Autonomous miles Reportable Disengages Disengagements per 1,000 miles Google, 2015 424,331 341 0.80 Google, 2016 635,868 124 0.20 BMW, 2016 638 1 1.57 Mercedes-Benz, 2016 673 336 498.95 GM Cruise, 2016 9,730 149 15.31 (b) (a)

36 Advanced Ground Vehicle Technologies for Airside Operations can be used to jam GPS in AGVT to create false locations and confuse the internal systems. Many AGVT use redundant GPS systems to determine their location and route of travel; however, these have been identified as weak points where hackers can disrupt operations. The U.S. Government Accountability Office (GAO) also did a study in cybersecurity vulner- abilities in vehicles, but focused on how NHTSA needs to redefine its role in attacks. The study found that NHTSA has not formally defined their roles and responsibilities in the event of a real vehicle cyberattack. If an attack were to occur, the response would be slowed since procedures are not in place (U.S. GAO, 2016a). The GAO report also outlines how hackers could attack CVs, potentially disabling vehicle brakes through a phone that is connected via Bluetooth to the onboard computers (U.S. GAO, 2016a). Figure 14 from the GAO report shows potential mitigation systems for these cybersecurity weaknesses. Airports need to consider cybersecurity for AGVT just as they need to consider cybersecurity in other airport activities. Perhaps AGVT cybersecurity issues can serve as a catalyst for more comprehensive airport cybersecurity programs. Testimony in front of a U.S. House of Repre- sentatives panel in October 2018 acknowledged that as recently as 2015, only 34% of airports responding to an ACRP survey had implemented a national cybersecurity standard or frame- work (Biesecker, 2018). Timeframe for Deployment As with many technologies, the timeline for AGVT implementation is often difficult to predict. Many companies have forecasted expectations regarding deployment (Table 4) but the timetable depends on a variety of factors. Tesla currently claims that the vehicles have the hardware needed for full self-driving. Tesla sales people tell consumers that it will be activated “as soon as it is legal” and Elon Musk said that a Tesla can drive from Los Angeles to New York by the end of 2018 without a human touching the steering wheel (Walker, 2018). This is consistent with Tesla statements in 2017 that L3 or L4 technologies will be ready for deploy- ment during 2018, although it has not been fully realized. Tesla vehicles receive regular software updates that are downloaded to the onboard computers, which directly impact automation capabilities (Tesla, 2018), a capability that presents challenges of its own, since drivers may not always be aware of changes to the automation. Waymo, the Google spinoff company that focuses on autonomous vehicles, planned to deploy a driverless car service in December 2018 for a group of Waymo volunteers near Phoenix, Arizona, who have participated in previous trials (Randall, 2018). In many cases, companies are stepping back on aggressive deployment with some AV companies adding humans back in the vehicle (although not necessarily in the Figure 14. Potential mitigation technologies to support in-vehicle cybersecurity. Image: GAO analysis of stakeholder information in GAO-16-350, as referenced in GAO, 2016a.

Applications and Lessons Learned 37 driver seat) and others having vehicles contact a remote operator to assist in determining the appropriate action. In 2018, Pronto engineer Anthony Levandowski reported a successful drive from coast-to-coast without a human touching the steering wheel. The vehicle was an autonomous prototype with L2 technology (termed Copilot) that required the driver to remain responsible for monitoring the environment. The technology was developed for class-eight semi-trucks with no expectation of replacing drivers (Glon, 2018). One challenge with predictions of deployment timetables is that, even if the technology is ready for the roadway environment, there may be modifications needed before it is ready for airside deployment because of differences in the operational context and the risks in the airside environment. For example, AV designed for the roadway may be calibrated to recognize ground-based obstacles such as people or other vehicles, rather than obstacles above, such as aircraft wings. Automation Research in Transit and Roadway Sectors In addition to considering deployment in the roadway sector by private firms, it is useful to consider research plans for deployment in the transit and roadway sectors. In 2018, FTA released a Strategic Transit Automation Research Plan. This plan includes a 5-year research roadmap; selected projects are shown in Figure 15. The research roadmap identifies enabling research for automation, followed by integrated demonstrations and strategic partnerships with both the private and public sector. Strategic partnerships will not only leverage expertise Organization 2017 Forecast (Lewis, Roger, and Turner, 2017) 2018 Forecast (Walker, 2018) Forecasted Year of Deployment Automation Level Forecasted Year of Deployment Automation Level Ford Motor Company 2021 L4 2021 L4 Certain cities Uber 2021 Not specified Volvo 2021 L4 2021 L4 Highway General Motors 2020 Not specified Tesla 2018 L3 or L4 2018 L4 Google/Waymo 2020 L4 Honda 2020 L4 Highway Toyota 2020 L4 Highway Renault-Nissan 2020 L4 Urban 2025 L5 Hyundai 2020 L4Highway 2030 Urban Daimler 2020s L4 and L5 Urban National Association of City Transportation Officials 2020 L4 IHS Markit (Market Research Company) 2020 L4 and L5 ABI Research (Market Research Company) 2021 L4 and L5 Table 4. Expected year of deployment for L3, L4, and L5 technologies.

Figure 15. Selected projects in strategic transit automation research roadmap. Based on information in FTA, 2018.

Applications and Lessons Learned 39 from partners, but will also enable the FTA to access datasets. These strategic alliances will facilitate progress and results that would be unavailable otherwise. FTA identified automation as a topic of interest over a decade ago, which led to the devel- opment of several research projects such as Vehicle Assist and Automation (VAA), a seven-year project (2009 to 2016) that installed VAA on a 60-foot bus and used automation to precisely dock at bus stops and provide lateral control in narrow lanes. Other AV projects included a $4.2 million project in which the Minnesota Valley Transit Authority developed their own automated bus technology (it started in 2008) and support of the CalTrans Partners for Advanced Transportation Technology (PATH) project, which began conducting Intelligent Transportation Systems (ITS) research in collaboration with the University of California, Berkeley, in 1986 (FTA, 2018). In the roadway sector, NCHRP funded project NCHRP 20-24(98) to develop a roadmap for research to support AVs and CVs in the roadway sector. This research was completed in 2015 and identified a list of unresolved issues related to the following categories: institutional and policy, operational, legal, and planning. The project also identified seven institutional and policy projects, 10 infrastructure design and operations projects, three transportation plan- ning projects, and three transit and freight projects (termed modal applications), totaling over $15 million proposed for investigation over a 5-year period. Some of these projects have already been completed and are underway, such as NCHRP 20-102, Impacts of Connected Vehicles and Automated Vehicles on State and Local Transportation Agen- cies, which has components that were completed in 2017, and other components that will be under- way through 2019, with 24 tasks identified which will be supported by $6.5 million in funds. A list of the tasks that have been identified provides an appreciation for the range and breadth of issues: • Policy and planning actions to internalize societal impacts of CV and AV systems into market decisions, • Impacts of regulations and policies on CV/AV technology introduction in transit operations, • Challenges to CV and AV application in truck freight operations, • Strategic Communications Plan for NCHRP Project 20-102, • Road markings for machine vision, • Implications of automation for motor vehicle codes, • Dedicating lanes for priority or exclusive use by CVs and AVs, • Updating regional transportation planning and modeling tools to address impacts of CVs and AVs, • Cybersecurity implications of CV/AV technologies on state and local transportation agencies, • Mobility-on-demand and automated driving systems: a framework for public-sector assessment, • Business models to facilitate the deployment of CV infrastructure to support AV operations, • Planning data needs and collection techniques for CV/AV applications, • Data management strategies for CV/AV applications for operations, • Impacts of CV/AV technologies on highway infrastructure, • Preparing TIM responders for CVs and AVs, • Deployment guidance for CV applications in the open source application development portal, • Minimum safety data needed for AV operations and crash analysis, • Update AASHTO’s CV/AV research roadmap, • Infrastructure modifications to improve the operational domain of AVs, • State and local impacts of automated freight transportation systems, and • Infrastructure enablers for CVs, AVs, and shared mobility in the near-term and mid-term. Other projects have been identified but have not yet begun. For example, NCHRP RFP 14-42, Determining the Impact of Connected and Automated Vehicle Technology on State DOT Main- tenance Programs, is expected to begin in 2019. Projects in the NCHRP research roadmap are shown in Figure 16.

40 Advanced Ground Vehicle Technologies for Airside Operations Figure 16. Projects in NCHRP CV/AV research roadmap. Based on information in Shladover, n.d.

Next: Chapter 4 - Prioritized Airside Applications »
Advanced Ground Vehicle Technologies for Airside Operations Get This Book
×
 Advanced Ground Vehicle Technologies for Airside Operations
Buy Paperback | $87.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Recent advancements in automated and advanced driving technologies have demonstrated improvements in safety, ease and accessibility, and efficiency in road transportation. There has also been a reduction in costs in these technologies that can now be adapted into the airport environment.

The TRB Airport Cooperative Research Program's ACRP Research Report 219: Advanced Ground Vehicle Technologies for Airside Operations identifies potential advanced ground vehicle technologies (AGVT) for application on the airside.

Appendices B Through S are online only. Appendix A, on enabling technologies, is included within the report.

READ FREE ONLINE

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!