51��Ƶ

The big national grants just keep rolling in for 51��Ƶ-Dearborn researchers. One of the latest is poised to have a broad impact across the engineering and computer science programs. According to the grant summary, a team of five faculty have been approved to “acquire an Autonomous Plug-In Hybrid Vehicle research platform to support collaborative and multidisciplinary research and education activities in several areas, including embodied cognitive vehicle, in-vehicular network security, energy consumption, environmental perception, cybersecurity, and driver behavior analyses in electric and advanced mobilities.” Translation: We’re getting funding to buy a fully programmable, road legal autonomous vehicle that will enable our faculty and student researchers to do all kinds of things they’ve never been able to do before.

Assistant Professor of Electrical and Computer Engineering (ECE) and principal investigator Jaerock Kwon says they hope to finalize their order for a specially equipped Chrysler minivan from Rochester Hills-based in the coming months. That could put the vehicle on campus later this year or early next year. The van will be loaded with a full suite of sensors needed for fully autonomous driving, including advanced optical cameras and a system for creating 360-degree maps of the surrounding environment. One of the things that’s cool about this particular vehicle is that it won’t require a bunch of clunky hardware add-ons. Kwon says typically when researchers adapt standard model vehicles for autonomous driving, they have to install third-party actuators to mimic a driver’s physical interactions with the car (e.g. turning the steering wheel, or depressing the brake or accelerator pedals). This vehicle doesn’t need any of that. Instead, it will use a “drive by wire” kit, which will allow commands to be sent directly to the vehicle's internal computer systems. From a mechanical and programming perspective, it’s much more elegant. Plus, it more closely approximates the likely anatomy of the AVs that will come to market in the future.

Across the board, the faculty leading this project say the AV is going to give them a chance to finally field test solutions they’ve been working on in their labs for years. “For a lot of the work we do, we can build models and simulations to test our hypotheses, but when we reach the question of how we’re going to actually test something in a real-world setting, that’s always a challenge,” says Hafiz Malik, an ECE professor who specializes in the cybersecurity of cyber-physical systems. “Getting this kind of equipment will enable us, for example, to validate how certain attack vectors or countermeasures play out on real vehicles.” Beyond cybersecurity, which will be a focus for Malik and ECE Assistant Professor Alireza Mohammadi, there are other research areas the vehicle could be a boon for. For example, Assistant Professor of Mechanical Engineering Youngki Kim plans to test drive algorithms that improve energy efficiency and the user experience. ECE Associate Professor Samir Rawashdeh, a robotics perception specialist, says the vehicle's arsenal of sensors could be extremely helpful for generating the vast amounts of data necessary for training and fine tuning machine learning algorithms. Meanwhile, Kwon, whose background is in brain science, is planning on exploring a wild new frontier in autonomous vehicle research. He thinks bigger breakthroughs might lie ahead if AVs were designed to function more like “cognitive beings” rather than giant computers. (We’ll have more on Kwon’s “embodied vehicle” research in a story later this fall.)

Beyond specific projects, the research vehicle could have cascading effects for the university. For one, Malik says having an advanced instrument that enables real-world validation of ideas can make future 51��Ƶ-Dearborn proposals much more attractive to large funders or industry partners, who are generally drawn to more implementation-ready solutions. It could also be a big boost for our student researchers. Malik says someone who can demonstrate knowledge of algorithms and simulations is great. But a student who can show a prospective employer that the solutions they helped develop were literally ready for the road may be hard to beat.

###

This 51��Ƶ-Dearborn project is being supported by the National Science Foundation , which enables the development or acquisition of multi-user research instruments that are critical to the advancement of science and engineering.