Abstract
This paper presents a method of evaluation of reliability of particle used for the particle filter using gaussian mixture model (GMM) which is made from the relation of the the observation obtained from the sensor and the state that is the presumption object. The particle filter is known to be effective for tracking people. However, when we evaluate reliability of particle, particle filter often needs a complex, nonlinear transformation between the amount of the state and the observation. Then, a proposal method is evaluation reliability of particle directly from the GMM. So, there are not needed a complex, nonlinear transformation. Moreover, we propose a particle filter with multiple prediciton models wherein the prediction model is selected according to the likelihood value of each prediction model. Furthermore, the proposed method is applied to the carrying task using mobile robot and the effectiveness is verified.
