Monocular Vision Tracking Based on Hybrid Particle Filters for a Person Following Robot

Abstract

A person following behaviour for a mobile robot with a new vision tracking algorithm is presented in this paper. According to the different characteristics of particle filter and Kalman filter, a novel approach of target tracking based on hybrid particle filters is applied to process the target object's position and shape component respectively, whose state updating is on the basis of data fusion between these two filter algorithms. The proposed method can not only pave the way for a low-complexity particle filter algorithm in dealing with higher dimensional tracking problem, but also cover the shortage of Gaussian restriction in Kalman filter. With the result of hybrid particle filter and the projective model of camera, the distance between the target and the robot can be calculated in real-time so that the robot can decide its own action to follow the target autonomously. A series of experiments on the Pioneer 3 robot show the method's validity and practicability.