Frontiers of Engineering Reports on Leading-Edge Engineering from the 2012 Symposium (2013) / Chapter Skim
Currently Skimming:
The Car and the Cloud: Automotive Architectures for 2020--Rahul Mangharam
The Car and the Cloud: Automotive Architectures for 2020--Rahul Mangharam
Pages 77-92
The Chapter Skim interface presents what we've algorithmically identified as the most significant single chunk of text within every page in the chapter.
Select key terms on the right to highlight them within pages of the chapter.
From page 77... ...
To that end, our automotive research team at the University of Pennsylvania is developing an in-vehicle programmable system, AutoPlug, an automotive architecture for remote diagnostics, testing, and code updates for dispatch from a datacenter to vehicle electronic controller units. For connected vehicles, we are implementing a networked vehicle platform, GrooveNet, that allows communication between real and simulated vehicles to evaluate the feasibility and application of vehicle-to-vehicle (V2V)
|
From page 78... ...
platforms. As new automotive electronic architectures are developed to enable remote diagnosis and reprogrammability throughout the life of the vehicle, drivers will be able to choose from a software component marketplace to enhance the safety, performance, and comfort of their vehicle.
|
From page 79... ...
IN-VEHICLE SYSTEMS: REMOTE DIAGNOSTICS, TESTING, AND REPROGRAMMING More than 20.3 million vehicles were recalled in 2010, many because of software issues related to electronic systems such as cruise control, antilock braking, traction control, and stability control. New and scalable methods are necessary to evaluate such controls in a realistic and open setting.
|
From page 80... ...
7 Current automotive systems lack a systematic approach and infrastructure to support postmarket runtime diagnostics for control software (although at least one online source indicates that there is a significant effort to incorporate automotive software testing and verification at the design stage8)
|
From page 81... ...
We have developed an early design of such a system, AutoPlug, although we recognize that the approach will be difficult in practice because it would require extensive runtime verification of the updated controller. Overview of AutoPlug AutoPlug is an automotive ECU architecture between the vehicle and an RDC to diagnose, test, update, and verify control software.
|
From page 82... ...
The framework integrates in-vehicle and remote diagnostics and makes vehicle warranty management more cost-effective. The aim of the AutoPlug architecture, illustrated in Figure 2, is to make the vehicle recall process less reactive with a runtime system for diagnosis of automotive control systems and software.
|
From page 83... ...
sends custom diagnostic code to the vehicle to observe its performance. Using vehicle models developed during the design phase, the RDC safely observes the operation of the software on the vehicle while it is running.
|
From page 84... ...
, traction control, cruise control, and stability control to see that the testbed does indeed perform as a real vehicle would. The main contributions of our applied research and development are threefold: • an architecture that uses both in-vehicle and remote diagnostics for remote recall management of deployed vehicles; • modification of the traditional observer-based fault detection and isola tion scheme for in-vehicle opportunistic diagnosis, as well as an experi mental thresholding scheme in the presence of modeling uncertainties; and • implementation and evaluation of these schemes on real ECUs for hardware-in-loop simulation.
|
From page 85... ...
of a particular sensor and its expected values. A smart thresholding scheme is used at the remote diagnostics center to determine the extent of the fault based on the residual signal.
|
From page 86... ...
With connected vehicles, it is necessary to analyze and validate the effect of incremental deployment of V2V technologies on message delay, coverage, and persistence in the region of interest. Because it is expensive to develop and test experimental protocols on a large fleet of vehicles, there is a need for vehicular network simulators that faithfully model first-order effects of the street topology, vehicle congestion, speed limits, communication channels, and spatiotemporal trends in traffic intensity on the performance and reliability of V2V networking.
|
From page 87... ...
For this test we drove five real vehicles along Forbes Avenue in Pittsburgh and conducted experiments with more than 4,000 virtual vehicles. Such hybrid simulation provides application users with an intuitive feel of the impact of communicating vehicle density on packet delivery ratio and event response time, and provides the developer with feedback about accuracy and details needed in the simulation models.
|
From page 88... ...
88 FIGURE 4 Mixed evaluation of real and virtual connected vehicles with the GrooveNet platform. The three vehicles in the circles are real vehicles communicating with short-range wireless communication (using the IEEE 802.11p/WAVE protocol)
|
From page 89... ...
TRAFFIC CONGESTION ANALYSIS To better understand empirical models of traffic congestion in different street topologies across the nation, and to develop sound traffic prediction and congestion-aware fastest-path routing algorithms, it is necessary to analyze largescale traffic mechanisms. We have developed a traffic analysis tool, AutoMatrix, that simulates and routes over 16 million vehicles on any US street map and provides real-time traffic routing services with hierarchical and synthetic traffic matrices (Figure 6)
|
From page 90... ...
(Left) Traffic congestion simulation showing more than 800,000 vehicles in Washington, DC.
|
From page 91... ...
Using these approaches, AutoMatrix has the potential to improve response time to traffic incidents by advising drivers to take the updated fastest path to their destination. We are working to use live traffic congestion data to support the needs of urban transportation operation centers.
|
Key Terms
This material may be derived from roughly machine-read images, and so is provided only to facilitate research.
More
information on Chapter Skim is available.