Optimal network resource allocation for monitoring continuous and hybrid systems

Abstract

The goal of the proposed project is to apply topological entropy concepts and techniques fornonlinear continuous and hybrid systems to the problems of state estimation and model detectionover nite-data-rate communication links. We consider a scenario where an output function withnitely many values, such as a quantizer, provides measurements of the state of a continuous-timeprocess. First, we introduce a fundamental quantity called estimation entropy that correspondsto the minimal bit rate needed for estimating dynamical systems. We will develop upper andlower bounds on this quantity in terms of some contraction metrics in the general nonlinearcase, or matrix eigenvalues in the linear case. Next, we address the problem of estimating thestate of the continuous dynamical systems with desired eventual accuracy using coarse outputmeasurements. When sucient data rate is available, a concrete algorithm for state estimation ofcontinuous systems will be constructed. We will study the eciency gaps between the algorithmsand the lower bounds. We will then develop algorithms for state estimation of switched andhybrid systems which will require us to tackle the intermediate problem of determining which oneamong a set of modes (or continuous models) is operational. Finally, we will explore applicationsof these algorithms in the context of the automotive and aerial vehicle control and monitoring.

Document Details

Document Type
DoD Grant Award
Publication Date
Apr 09, 2018
Source ID
FA95501710236

Entities

People

  • Sayan Mitra

Organizations

  • Air Force Office of Scientific Research
  • United States Air Force
  • University of Illinois Urbana–Champaign

Tags

Fields of Study

  • Computer science

Readers

  • Adaptive Control and Estimation with Uncertainty in Dynamic Systems.
  • Distributed Systems and Data Platform Development