An Entropy Framework for Communication and Control in Cyber-Physical Systems

Abstract:

Many practical systems can be categorized as cyber physical systems (CPSs), such as smart grid, unmanned aerial vehicles (UAVs) and robotic networks. A typical CPS consists of physical dynamics, sensors, communication network and controllers. The communication network is of key importance in CPS, since it mimics the nerve system in the human body. Hence, it is critical to study how the communication network in CPS should be analyzed and designed. Essentially, communications stem from the uncertainty of system under consideration; meanwhile, random perturbations increase the system uncertainty, which will be reduced by the control actions in CPS. It is well known that entropy is a measure of the system uncertainty. A unified framework of entropy is used for CPSs, in which random perturbations create entropy while communications and controls provide negative entropy to compensate the entropy generation. The intellectual merits are the novel framework of entropy for bridging the communications and control in CPS and the new design criterion based on the entropy of system state for CPS. The project’s broader significance and importance are the education of various levels of students, the dissemination of results to public, and the impact on everyday life (such as to improve the agility and robustness of power grids). This project applies the framework of entropy to study the interdependencies of communications and control, thus facilitating the analysis and design of communications in CPSs. The following tasks are tackled in the project. Task 1. Entropy Flow Based Communication Capacity Analysis: In this task, the requirement of communications in CPS is analyzed by studying the entropy fields in the physical dynamics, thus providing an estimation on the scale (bits/second) of communication capacity budget. Task 2. Communication Network Topology Design: Based on Task 1, the design of the network topology (either physical or logical) in the realization stage is tackled in both optimization based or heuristic approaches. Task 3. Online Network Resource Scheduling: Based on the network topology obtained in Task 2, it refines the network resource scheduling during the operation, again, using both optimization based and heuristic approaches, within the framework of entropy fields. Task 4. Hardware Emulation Testbed: A co-simulation testbed is delivered, which is based on real time digital power simulator (RTDS) and professional communication simulator, in the context of smart grids. Based on the research, new courses are developed. K-12 outreach and various levels of undergraduate/graduate educations are incorporated into the research.