GOALI: Hybrid Dynamic Feedback to Design Provably Correct Driving

点击翻译按钮获取中文摘要

The research objective of this award is to establish fundamental theory, develop computational techniques, and prototype real-world implementations for next generation driver-assist systems that ensure safety and fuel efficiency at traffic intersections. The research will result in algorithms for smart vehicles that cooperate with each other at intersections, merges, and stop signs to guide their drivers toward provably safe (no collisions) and fuel efficient driving choices. The research approach is to employ hybrid dynamical models (systems described by differential equations and logic) with imperfect information and solve a safety control problem for hybrid systems with unknown state. The industry partner will enable the development of prototypes of cooperative active safety systems that ensure collision-free and fuel efficient driving. Deliverables include algorithms and prototypes for cooperative active safety systems that account for fuel efficiency.If successful, the results of this research will allow the development of cost-effective and guaranteed driver-assist systems for safety and efficiency. Therefore, it has the potential of impacting society by lowering American roadway fatalities. The driver-assist systems for fuel efficiency will contribute in lowering the economic cost of vehicle driving and reduce emission, with clear benefits to the environment. This work promotes a synergy of sensing, communication, and control technologies. This along with the industry collaboration will potentially bring a substantial technological advancement to the automotive industry in the near future. Students working on this project will have the unique opportunity to work on cutting-edge technology at the industrial site collaborating with expert practitioners. The research results will be disseminated broadly through various means, including workshops at national and international meetings and more focused workshops at both the industry site and at MIT.