Adaptive Estimation of a Moving Object Trajectory Using Sequential Hypothesis Testing

Abstract

The present paper addresses the problem of adaptive estimation of a moving object trajectory and detection of changes in the motion mode. It is supposed that an object moves along a complex trajectory and at known discrete-time instants it may change its motion to one of three possible modes: a uniform straight line motion or a uniform anticlockwise/clockwise circular motion. We propose a new algorithm for adaptive trajectory estimation that combines a hybrid linear stochastic model of an object trajectory with a bank of competitive Kalman filters and a decision rule based on a sequential hypothesis testing. A detailed description of the decision rule and pseudocode of the proposed algorithm are given. The software implementation of the algorithm is made in Matlab. A numerical example of adaptive estimation of the motion of an object along a complex trajectory consisting of nine different pieces is considered. We have conducted computational experiments with different levels of noise in the measurements. The results confirm the effectiveness of the proposed algorithm