A team consisting of Bucknell professor Craig Beal (mechanical engineering), Penn State professor Sean Brennan (mechanical engineering), and University of Massachussets Lowell professors Cindy Chen (computer science) and Kshitij Jerath (mechanical engineering) have been given a $1.2 million National Science Foundation grant to study and use data from connected "intelligent vehicles" that could lead to safer and more ubiquitous self-driving vehicle technology.
Their studies focus on using the data to program self-driving vehicles to account for and navigate hazardous road conditions such as icy patches.
“This work can save lives and suffering by improving safety-critical decision-making used by connected and autonomous vehicles driving under hazardous operational situations,” Beal said. “This research is timely and impactful due to the recent and rapid deployment of vehicles with enhanced sensing and connectivity."
Beyond this, Beal explains, some of the data that the vehicles collect can muddle the computers' decision-making algorithms: “This grant is really intended to investigate how data points are related to each other and when data are no longer useful and actually lead to worse computer decision-making. It has broader applications beyond vehicles, such as in predicting yields of crops in fields planted at specific times within a region, or determining for how long Yelp user reviews of a restaurant provide useful information.”
Professor Beal previously received a 2017 NSF grant to measure steering torque — the effort required to hold the front wheels at a given angle as a vehicle travels around a turn — through sensors mounted within the wheels of an autonomous vehicle. The new grant builds upon that work to study how data collected through steering sensors in a fleet of connected and autonomous vehicles can be shared in a timely fashion to improve driving safety under hazardous road conditions.
For example, when several connected vehicles hit a slick patch on an icy roadway, they could alert other connected vehicles in the area to take corrective action.
Beal hopes that the methods and algorithms developed in this work can guide the deployment of all database systems where trust in data quality and timeliness are competing factors.
“This enables a wide range of information sharing,” he said. “The crux of what we’re trying to study is how you can use data to make the best decision. In a vehicle application, that includes where are you going, how fast should you go, and what the conditions ahead are like. We’re using the vehicle as an important and socially acceptable example, but there are a lot of examples that are becoming data-connected systems.”
The grant also includes funding for outreach activities to provide young women with increased knowledge and role models, with the aim of increasing their participation in study of the engineering and computing fields. It will also support the development of graduate and undergraduate researchers at Bucknell, Penn State, and UMass Lowell. Bucknell undergraduate students will be participating in the research project.