Stable Target Tracking using Observer Based Velocity Estimation

Abstract

In applications where mobile robots are used to track non-cooperative moving objects it is often required that not only the position but also the velocity of the moving target can be measured. In this paper, we consider a problem in 2D where the tracking robots are equipped only with vision and position sensors and are unable to measure target velocity directly. Instead, two separate observers for target velocity are proposed and shown to stabilize the two tracking controls used by the robots. To evaluate the observers, results from simulations with observer based velocity estimates are compared to corresponding results where the velocity estimates are given by the standard Extended Kalman Filter algorithm.