Location: Ravelijn RA 2501
Title: Universally optimal particle-filter-based detection (joint work with Pranab Mandal)
Abstract: In this talk we will address the problem of optimally detecting a signal or object. Or stated otherwise optimally deciding between two hypotheses about signal/object presence. We consider possibly nonlinear and nonGaussian signals. It is shown that we can construct a detector using quantities from a running particle filter. Furthermore we go into the optimality of the detector. We also provide a proof of the optimality using stochastic convergence. We illustrate our theoretical results with a simulation example.
Short Bio of Yvo Boers.
Yvo received the M.Sc. degree in Applied Mathematics at Twente University in 1994. His Thesis was on adaptive control under the guidance of Dr. Jan Willem Polderman. He received the Ph.D. degree from Technical University Eindhoven in 1999. His thesis was on robust controller design for systems with Statistical uncertainty. His main advisor was prof. dr. Siep Weiland. From 1999 until 2012 He worked at Thales, Hengelo. Mainly in the fields of object tracking, signal processing and particle filtering. From 2005-2012 Yvo also was a research associate in the systems and control group of Applied Mathematics of Twente University on a personal NWO grant. After 2012 Stroke Yvo retired from his regular jobs and does research as a hobby.
