Event-triggered Estimation: Observability, Identifiability and Parameter Learning

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering

Final Year Thesis Oral Defense

Title: "Event-triggered Estimation: Observability, Identifiability and 
Parameter Learning"

by

LI Kin Fung

Abstract:

Event-triggered state estimation is an emerging area in control theory. The 
extraction of information in event-triggered output measurements is crucial in 
analyzing system characteristics such as observability and identifiability. 
Meanwhile, the tools for observability analysis in other types of systems are 
well-developed. Qualitative tools include the observability matrix for linear 
systems and the observability Lie algebra for nonlinear systems. Recently, 
quantitative measures based on the empirical Gramian, such as the 
unobservability index and the estimation condition number, are defined on more 
general systems, possibly nonlinear and stochastic. This study proposes a 
quantitative extension of system observability to event-triggered systems 
through the empirical Gramian and examines its relevance to qualitative system 
properties. Practical considerations in the implementation of observability 
criteria and parameter learning algorithms for event-triggered systems are also 
discussed.


Date            : 3 May 2024 (Friday)

Time            : 09:45 - 10:25

Venue           : Room 6602 (near lifts 31/32), HKUST

Advisor         : Prof. CHAN Gary Shueng-Han

2nd Reader      : Prof. SHI Ling