Mobile Robot Localization Based on Multi-Sensor Model for Assistance to Displacement of People with Reduce Mobility

This paper deals multi-sensor data fusion problem for mobile robot localization. In this context, we have used data fusion sensors: encoders and ultrasonic sensor. To improve the robustness of localization and to reduce the estimation error we have proposed a Kalman Particle Kernel Filter (KPKF) approach, which is based on a hybrid Bayesian filter, combining both extended Kalman and particle filters. The KPKF filter using a Gaussian mixture in which each component has a small covariance matrix. The Kalman correction updates the weights in order to bring particles back into the most probable space area. This method can be applied for non-linear and multimodal environment and can improve localization performances and reduced estimation error. The proposed approach is implemented on a LIASD-Wheelchair experimental platform. Keywords—Localization; multi-sensor; data fusion; mobile robotics; Kalman filter; particle filter; smart wheelchair.