CPS: Small: MPSoC based Control and Scheduling Co-design for Battery Powered Cyber-Physical Systems

Abstract: The objective of this research is to develop new scientific and engineering principles, algorithms and models for the design of battery powered cyber-physical systems whose computational substrates include high-performance multiprocessor systems-on-chip. The approach is to design control tasks that guarantee performance and meet criteria for battery operation time. Task schedulers are co-designed to balance the computing load across the multiple processors, and to control the physical plant together with the control tasks. The controller and scheduler will be integrated with battery management algorithms through a systems theory approach so that the methods are provably correct with justfiable performance. Intellectual Merit: The program will create progress in digital and hybrid control theory that keeps up with the recent trend of using multiprocessor systems-on-chips for control and robotic applications. The mechanism for the migration of control tasks between multiple processors will respect physical and thermal performance. A novel battery dynamic discharge model is developed, which may be applied to context when the discharge current of batteries cannot be predicted by existing static battery models. Broader Impacts: Collaborations with industrial partners have been set up. The program offers multidisciplinary training in cyber-physical systems. A teaching and outreach lab is in place to host K-12 student teams that participate in robot competitions, and has become an Explorer Post for Boy Scout of America.