Strategic Issues Facing Transportation, Volume 4: Sustainability as an Organizing Principle for Transportation Agencies (2014)

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.

148 Today, the policy, governance, funding flow, and decisionmaking challenges of TBL sustain- ability are daunting. Some thought leaders in the transportation sector are uncomfortable with the implications of managing transportation within a sustainable society policy system. In the prior chapter, the research team identified significant challenges and development needs facing application of key models and “must have” analysis methods that decisionmakers and managers will certainly need to navigate in a policy system based on TBL sustainability as an organizing principle. Nevertheless agencies in the United States and around the world have undertaken many sustainability initiatives that are advancing sustainability objectives. Progress is being made, but it remains primarily at the programming and project delivery functional levels, focused on sustainable transportation. At this stage, the data and supportable models do not exist to link transportation investments to results at three bottom lines—nor to establish and communicate TBL strategy alternatives simply and clearly to the public and among high-level decisionmakers. However, there is good reason to believe that, via dramatic technological advances over the next 10 years, the data and TBL modeling challenges seen today will greatly diminish. This will dramatically transform the ability to understand and make the informed high-level TBL decisions with confidence, and with the support and acceptance of the public. The research team envisions major technological and scientific advances in four critical areas: • Improved sensor technology, rapid data capture, data mining, analysis, and processing of “big data” • Development of more powerful intelligent computing and “learning models” for decision making • Continued rapid advancement in interpersonal communications, networking capabilities, social media, and information dissemination to greatly facilitate engagement of the public • Much greater scientific and public understanding of the relationship between transportation and the economy, society, and the environment Each of these areas is discussed in the following sections. 9.1 Improvement in Sensor Technology and Omnipresent Data Collection “Big data” refers to the collection of data sets so large and complex that it becomes difficult to process using conventional database structures and data management tools. Every day modern sensor and data collection tools collect, manipulate, and store enormous quantities of data. Business, defense, energy management, meteorology, genomics, complex physics simulations, C H A P T E R 9 Addressing TBL Sustainability— Now and in the Not-Too-Distant Future

Addressing TBL Sustainability—Now and in the Not-Too-Distant Future 149 and biological and environmental research now regularly deal with petabytes to exabytes of data. The world’s technological per-capita capacity to store information has roughly doubled every 40 months since the 1980s such that now 2.5 quintillion (2.5 × 1018) bytes of data are created every day (IBM, n.d.). Furthermore, the rate of big data growth is accelerating. Of the total data existing in the world today, more than 90 percent has been created in the last 2 years. Data sets are growing in size in part because they are increasingly being gathered by ubiquitous information-sensing mobile devices, aerial and space-based sensory technologies (remote sensing), software logs, cameras, microphones, radio-frequency identification readers, and wireless sensor networks. In addition, the emergence of connected vehicles and smart infrastructure means that the transportation system is about to become a major supplier of big data. Data from connected vehicles alone will add exabytes of data to the data stream and transport the potential for managing the transportation system to reduce its environmental footprint. Table 44 shows some examples of these technologies, which are currently available or under development. As they are implemented, these technologies will radically increase the data available to transportation managers and increase managers’ ability to understand the impacts of vehicle operations (e.g., private automobiles, transit, and trucks) on the environment. In addition, a new generation of environmental sensors is emerging that will dramatically increase our understanding of how the environment is changing. Environmental sensor networks are rapidly evolving from passive logging systems that require manual downloading, into intelligent sensor networks that compose a network of automatic sensor nodes and communications systems which actively communicate their data to a sensor network server where this data can be integrated with Technology Description Eco-Approach and Departure at Signalized Intersections The Eco-Approach and Departure at Signalized Intersections application uses wireless data communications sent from roadside equipment to vehicles and encourages green approaches to signalized intersections. Eco-Traffic Signal Timing The Eco-Traffic Signal Timing application is similar to current adaptive traffic signal control systems; however, the application’s objective is explicitly to optimize traffic signals for the environment rather than the current adaptive systems’ objective. Eco-Traffic Signal Priority The Eco-Traffic Signal Priority application allows either transit or freight vehicles approaching a signalized intersection to request signal priority. Eco-Ramp Metering The Eco-Ramp Metering application determines the most environmentally efficient operation of traffic signals at freeway on-ramps to manage the rate of entering automobiles. Connected Eco-Driving The Connected Eco-Driving application provides customized real-time driving advice to drivers, allowing them to adjust behaviors to save fuel and reduce emissions. Multimodal Traveler Information The Multimodal Traveler Information application provides pre-trip and en route multimodal traveler information to encourage transportation choices with reduced environmental impacts. Dynamic Eco-Routing The Dynamic Eco-Routing application determines the most eco-friendly route, in terms of minimum fuel consumption or emissions, for individual travelers. This application also recommends routes that produce the fewest emissions or reduce fuel consumption based on historical, real-time, and predicted traffic and environmental data. Eco-Integrated Corridor Management Decision Support System The Eco-Integrated Corridor Management Decision Support System application involves using historical, real-time, and predicted traffic and environmental data on arterials, freeways, and transit systems to determine operational decisions that are environmentally beneficial to the corridor. The Eco-Integrated Corridor Management Decision Support System is a data-fusion system that collects information from various multimodal systems. Table 44. Connected-vehicle technologies that will have positive environmental impacts.

150 Sustainability as an Organizing Principle for Transportation Agencies other environmental data sets. The sensor nodes can be fixed or mobile and range in scale appro- priate to the environment being sensed. For example, large-scale single-function networks tend to use large single-purpose nodes to cover a wide geographical area. Localized multifunction sensor networks typically monitor a small area in more detail, often with wireless ad hoc systems. In the future, sensor networks will integrate these three elements (heterogeneous sensor networks). The emergence of the cloud and real-time integration and analysis means that this data can be analyzed in real time to identify environmental trends and events [see Hart and Martinez (2006)]. These networks are creating a revolution in earth sciences that is likely to transform understanding of the interactions between the human and natural world over the next decades. In addition, next generation sensors are emerging that may further enhance the ability to sense the environment. For example, the Defense Advanced Research Projects Agency (DARPA) is funding research into “smart dust,” a miniature wireless node that uses microelectromechanical sensors on a cubic millimeter scale [see Pister (1997)]. Billions of these micro sensors could be introduced into the environment to sense changes in environmental conditions and changes. Larger static sensors (approximately one centimeter long) could also be built specifically for an environment under investigation and embedded in the environment to track environmental changes [see Hart and Martinez (2006)]. In addition, big data can provide much more information on economic and social conditions. Currently national income accounts (the basis of any economic model) rely on data that is at least 3 months out of date. GDP growth rates, employment data, and price information are frequently in error and must be adjusted or corrected months after the first information is released. Real- time data offers the prospect of a more accurate, deeper understanding of economic impacts. Simultaneously new data collection and analysis methods related to social behavior offer the prospect of a better understanding of how people respond to specific changes in their environment (Madan et al., 2010). Taken together, these trends suggest that over the next 30 years, there is likely to be an explosion of data concerning transportation, economic, environmental, and social phenomena. This opens up the possibility of a better ability to understand, design, and operate transportation systems to optimize sustainability. One of the key requirements will be the development of processing capabilities that can analyze and understand the implications of this data. This issue is discussed in the next section. 9.2 Improved Computing Power, Advanced Analytics, and Intelligent Analysis and Modeling Along with increases in data, there has been a steady ongoing increase in processing power. Moore’s law, the observation that over the history of computing hardware, the number of tran- sistors on integrated circuits doubles approximately every 2 years, has meant that huge increases in computing power have occurred and continue to occur. While it is generally acknowledged that the limits of Moore’s law are being reached if silicon-based technology continues to be used, technologies such as optical computers, quantum computers, and DNA computing offer the possibility of continuing to expand into the future [see Peckham (2012)]. The futurist Ray Kurzweil agrees that by 2019 the current strategy of ever-finer silicon photo- lithography will have reached its maximum development but speculates that this does not mean the end of Moore’s law: “Moore’s law of integrated circuits was not the first, but the fifth, paradigm to forecast accelerating price– performance ratios. Computing devices have been consistently multiplying in power (per unit of time) from the mechanical calculating devices used in the 1890 U.S. Census, to [Newman’s] relay-based ‘[Heath]

Addressing TBL Sustainability—Now and in the Not-Too-Distant Future 151 Robinson’ machine that cracked the Lorenz cipher, to the CBS vacuum tube computer that predicted the election of Eisenhower, to the transistor-based machines used in the first space launches, to the integrated- circuit-based personal computer” (Kurzweil, 2001b). Kurzweil speculates that, given past performance and the demand for expanded data processes in the emerging era of big data, a replacement technology will arrive that will continue Moore’s law long after 2020. More hardware does not by itself provide machine intelligence. However, it contributes to the development of machine intelligence. For example, cheap, fast computing resources allow researchers to experiment with new algorithms and data mining and correlation techniques to achieve improved machine intelligence or to gain new insights by processing massive data sets (Muehlhauser and Salamon, 2012). There have been recent breakthroughs in artificial intelligence (AI; e.g., the development of Gödel machine formulations and AIXI11), and the near future promises significant breakthroughs in everyday application of AI (e.g., natural language recognition and translation, visual recognition, automated decisionmaking), massive data sets, progress in cognitive science and neuroscience, accelerating scientific development, and substantial economic incentives (Muehlhauser and Salamon, 2012). All of these developments suggest that conceivably human-level AI could be approached within the next 50 to 100 years. Moreover the development may lead quickly to machine super-intelligence (or so-called “strong AI”), that is, an intelligence that surpasses human intelligence and is capable of expanding its own intelligence in the future. Such systems would have dramatic effects on the ability to understand and manage the inter- action between transportation and the TBL. Combined with omnipresent data, it would be possible for such systems to immediately model and understand the relationship between the human and environmental system and determine the impacts of individual changes on the system as whole. This is not science fiction. Even now, the beginnings of such systems exist. Advanced analytic systems used in business and government combine multivariate statistical analysis, complex multistate simulation, machine learning, neural networks, text analytics, advanced data visualization, and other advanced data mining techniques to create near-intelligent decision supports systems. These systems can perform the following functions: • Analyze and explain complex hidden trends: Through the use of statistical analysis and modeling and simulation, advanced analytics provides a mechanism to dynamically analyze and understand trends and develop explanations for them. • Optimize operations and understand relationships: Advanced analytics helps analysts understand the relationships between variables, identify causality and system drivers, and optimize processes to increase program performance and goal attainment. • Predict outcomes and identify risks: By applying their understanding of past trends and the relationships between variables, analysts are able to develop a predictive model to iden- tify potential outcomes in real time and create a risk-based, probabilistic forecast of future states. 11 AIXI is the first mathematical theory of optimal Universal Artificial Intelligence. It was developed in 2002 and represents a major breakthrough in understanding automated decisionmaking that solves the problem of sequence prediction for un- known prior distribution. A Gödel machine is a mathematical formulation of a rigorous, general, fully self-referential, self- improving, optimally efficient problem-solving system. The model allows for a set of self-referential formulas that rewrites any part of its own code as soon as it has found a proof that the rewrite is useful, where the problem-dependent utility func- tion, the hardware, and the entire initial code are described by axioms encoded in an initial proof searcher which is also part of the initial code. The searcher systematically and efficiently tests computable proof techniques (programs whose outputs are proofs) until it finds a provably useful, computable self-rewrite.

152 Sustainability as an Organizing Principle for Transportation Agencies • Test assumptions, proposed policies, and alternative program models: Advanced analytics allows the creation of virtual worlds where assumptions concerning an analysis can be tested and the limitations and weaknesses of current models can be understood. In addition, it provides a sandbox where analysts can test proposed policies and/or alternative program models and assess impacts. • Visualize data: Advanced analytics allows previous numeric or quantitative data to be turned into a meaningful visual representation of the underlying order and regularity of the data. The combination of these tools and big data brings very close the kind of analytics required to properly understand and model the relationship between transportation and its impacts on the TBL. Given the development of AI and big data, it is quite possible that over the next decades there will be an explosion in the ability of automated systems to understand and manage sustainability. The cost of such systems (considering the continuing decline in technology and storage costs) may be such that the ability to manage sustainability in the future would be in the hands of almost every transportation agency, partner agencies, and the public. 9.3 Rapid Advancement in Networking Capabilities and Information Dissemination Leading to Improved Understanding of Relationships between Transportation and the Economy, Society, and the Environment While increased data and improvements in machine intelligence and analytics may provide the potential to better understand the impacts of transportation and other human systems on sustainability, a strong theoretical basis to model these interactions is still lacking. For example, economics is one of the most developed of the social and behavioral sciences and, due to the money that is associated with financial modeling, has attracted considerable resources in recent years. However, understanding of economic behavior is extremely limited. Macroeconomic models rarely predict accurately and are dependent on multiple assumptions and preprocessed data analysis. New approaches such as agent-based, system-dynamic models offer hope of improving their predictive capability but basic mathematical and theoretical challenges still exist. There is some hope that these challenges may be resolved. For example, the initiative known as the “Manhattan Projects for Economics” brings together physicists, mathematicians, and economists to work on extremely hard theoretical and computational problems in economics. The idea is that, because many of the challenges that economists face are similar to those that physical scientists have already encountered and resolved, the mathematical techniques and theoretical paradigms they have developed in resolving these challenges may provide insight into critical economic problems that economists are wrestling with [see Brown et al. (2008)]. For example, certain economic problems relating to competing multiple preferences may now be dealt with by infinite dimensional diff (S1) theory or Non-Abelian Gauge theory. The emergence of dramatically faster scientific communication via the web, huge increases in the number of scientists and research as China and India develop, and the resources that are already available worldwide suggest that this synergy and interdisciplinary communication may lead to breakthroughs in multiple fields over the next 10 to 30 years. Combined with improvements in AI, processing power, and the availability of data, it is very likely that current TBL decision- support challenges may vanish. With that, direct and supportable understanding of the relationship between transportation and TBL sustainability will provide the tools needed to manage resources in

Addressing TBL Sustainability—Now and in the Not-Too-Distant Future 153 real time to optimize and improve TBL sustainability to the maximum practical extent, should the policy system call for that goal. * * * * Whatever the future may bring, one of the limiting factors will still be the challenge we have as a society in reaching wide consensus on the relative value of the three bottom lines. While we may understand better how transportation affects the environment or how managing resources today affects the viability of the future, debate on those values will continue. No matter how the bottom line values are set and what predictive models and machine intelligence may tell us of the future—we will always encounter societal decisions that require us to assess how much we value the well-being of our next generation versus our own demands for quality of life.