Passivity-Based Target Tracking Robust to Intermittent Measurements

Abstract

A passivity-based switched systems analysis of an estimator-predictor framework is presented for mobile target tracking. Measurements of the target are provided by a mobile sensor with limited field-of-view (e.g., a camera). The analysis demonstrates that passivity-based dwell-times exist to ensure the tracking objective is achieved, despite intermittent feedback of the target pose. Specifically, a maximum ratio is determined between the length of time the target is unobserved and the length of time the target is observed. After a period of time where the target is unobserved, this provides a minimum time that the target must be observed to ensure passivity of the system. An additional result demonstrates conditions under which the estimator-predictor framework is uniformly ultimately bounded. These results show that the framework is robust to intermittent feedback, which is an alternative approach to relaxing the continuous observation constraint required by many existing results.