Experiment of a low cost computation particle filter for localization of an electric wheelchair

In this paper, an experiment of a novel particle filter for localization of an electric wheelchair is presented. The particle filter algorithm is modified in the sense of using a minimum number of particles, thus a low cost computation particle filter is performed. The prediction is calculated based upon the dead reckoning methods; the estimation is performed using the vision system which provides position and orientation measurements in the global coordinate. Only a small number of particles are used in the prediction step. The experimental results in the electric wheelchair show the effectiveness of the proposed method.