Wenxin Xiong

Computer Networks and Telematics, University of Freiburg

Tuesday, February 08, 2022, 11:00 - 11:59

"Online"

Source localization refers to determining the position of a signal-emitting target using sensor-collected measurements. The task turns out to be quite challenging when the robust location estimators are considered towards a better fit for the possibly erroneous signal observations in practice. This talk presents a series of efficient iterative algorithms to handle the non-convexity and/or nonsmoothness involved in the outlier-resistant positioning formulations. Here, we will elaborate on the specific cases of time-of-arrival (TOA), time-difference-of-arrival (TDOA), and angle-of-arrival (AOA) based localization, respectively, showing that different measurement models may lead to algorithmic designs with different emphases. The problem-solving paradigms being relied on range broadly from the alternating direction method of multipliers and Lagrange-type neurodynamic approaches to the half-quadratic technique and random search schemes. The main goal of this talk is to clarify the rich interface between several recently fashionable optimization-based iterative solutions and robust estimation problems arising in localization.