Distributed and Robust Estimation for Cyberphysical Systems using a Nonlinear Consensus Approach

Arizona State University, Scottsdale AZ

Investigators

Abstract

Objective: The objective of this project is to design and test nonlinear, robust, and power-aware consensus estimation algorithms for fully distributed cyber-physical systems (CPS). Systems that use a large number of networked sensors with limited transmission range due to power limitations are considered. Such systems have a large number of transmitters that are subject to non-Gaussian impulsive noise and network interference. Robust decentralized distributed sensing will be considered with joint design of the sensing and communication systems with low-power constraints. Intellectual Merit: The intellectual merit is in the novel design of distributed nonlinear consensus algorithms to solve parameter estimation problems for CPS applications. Proposed convergence analysis, and variance of the estimators in the presence of these nonlinearities go beyond existing analyses of linear approaches. The research will provide power-aware consensus estimation algorithms with minimal assumptions about the sensing and channel noise through judicious design of nonlinearities. Algorithms will be tested with photo-voltaic array data and a testbed will be developed to address and characterize real-time estimation. Broader Impact: The broader impacts are on alternative energy, environmental, infrastructure monitoring, and STEM education. The PIs will design educational software modules that use sensing as a paradigm to illustrate basic concepts in probability and communications at the undergraduate level. Dissemination and education efforts will include computer and mobile-based applications using the award-winning JAVA-DSP software packages. The investigators will also actively recruit students from under-represented minorities to work on the project.