This course focuses on teaching analytical and empirical tools for nonlinear network observability in autonomous systems. The learning outcomes include understanding how to determine which subset of nodes in a network should be measured to maximize information, investigating the effects of changes in network topology on observability, and extending observability metrics to nonlinear stochastic systems. The course covers skills such as sensor selection, optimal sensor placement, network observability, and mathematical modeling. The teaching method includes theoretical discussions, empirical demonstrations, and simulation exercises. The intended audience for this course includes engineers working on autonomous multiagent systems, researchers interested in network synthesis and disease spread tracking, and individuals involved in sensor placement for various applications.
Analytical and Empirical Tools for Nonlinear Network Observability in Autonomous Systems
Institute for Pure & Applied Mathematics (IPAM) via YouTube
Overview
Syllabus
Intro
Nonlinear Dynamics and Control Lab
Remote Sensing
Dynamics, Control, Sensing, Robustness
Agility and localization in biological systems
Active sensing in engineered systems: Wind-finding
Gyroscopic sensing in insect wings
Reduced-order modeling
Nonlinear observability
Observability via linearization about trajectory
Empirical observability Gramian
Limit case
Finite epsilon case
Fisher information bound
Sensor Selection - Problem framework
Sensor placement results
Optimal sensor placement
Network Observability
Optimization Algorithm
Virus Spreading Model (SIS)
Sparse or Dense Network Node Sensor Selection
Privacy in Networked Systems
Network Security
Mathematical Modeling
Optimal sensor locations for vortex sensing
Range-only and bearing-only navigation
Ongoing work
Acknowledgements
Taught by
Institute for Pure & Applied Mathematics (IPAM)
