Maneuver and Data Optimization for High Confidence Testing of Future Automotive Cyberphysical Systems

The current lack of toolchain for high confidence testing, validation and verification of advanced, connected and automated/autonomous vehicles can impede and even entirely prevent the introduction of such vehicles into mass production. To address this challenge, this projects develops theory, methods, and tools for generating and optimizing test trajectories and data inputs that can maximize opportunities to uncover faults in both physical and cyber domain in future automotive vehicles.  Our solution involves the following basic elements: (i) CPS/Smart Black Box with principled sampling-based vehicle data acquisition and management strategies to uncover faults in both existing and future vehicle fleets; (ii) Game theory-based simulation environment to inform in-traffic relevant trajectories; and (iii) Model-free trajectory optimization techniques for actively falsifying time domain specifications.  The ongoing research under this project advances CPS vehicle lifecycle management with focus on test generation and validation and verification. It also advances game theory, optimal control, information theory and data mining for applications in autonomous/automated vehicle setting. Data acquisition and sampling strategies being developed can be applied more broadly to connected vehicle & devices.   The ongoing research supports the automotive industry in introduction of automated/autonomous vehicle technology into mass production.  The automated/autonomous vehicles will have a significant societal impact, e.g. enabling transportation for people who are not able to drive.  Interdisciplinary advances being made are integrated into courses and tutorials and technology transfer and implementation opportunities are being considered through the collaboration with the industrial partner (AVL).