Self-Sustainable Data-Driven Systems In the Field

Abstract:

Data-driven intelligence is an essential foundation for physical systems in transportation safety and ef- ficiency, area surveillance and security, as well as environmental sustainability. While sophisticated data analysis and synthesis can be well supported in large data centers, future intelligent systems require on-the- scene processing with faster responses and less dependence on the unreliable (often wireless) data commu- nications in the field. Field processing must consume low power for easy deployment and self-sustainability (e.g., using solar energy harvesting and buffering). The combination of high-volume data processing with low-power computing and I/O creates critical challenges for future intelligent systems. This project will develop a new computer system infrastructure for the efficient utilization of low-power resources for innetwork data processing, for secure management of data in the field, and for task deploy- ment and migration. This project will also develop new image and data processing approaches for resource- adaptive fidelity adjustment and function partitioning. The synergy between the self-sustainable hardware, system software support, and application data processing will manifest through global coordination for quality-of-service, energy efficiency, and data privacy. Further synergy will be demonstrated through soft- ware exploitation and adaptation of self-sustainable hardware characteristics (e.g., specialized low-power computing and energy leakage of supercapacitors).