Using Sensor Fusion in a Monte-Carlo Localization System for Mobile Robot

摘要

In this paper, a method for multi-sensor system data fusion architecture is applied to a Monte-Carlo localization (MCL) system for an autonomous mobile robot in a known environment. This method uses particle filter to estimate the pose of robot. Then uses Kalman filter to reduce errors caused by wheel drifting. For this purpose, the accelerated information from Inertial Measurement Unit (IMU) corrects the encoder data from motor by Kalman filter. After that, according to the reliable information from odometry and vision data from omnidirectional camera, we can get the robot's position by using particle filter. This method not only reduces the error but also improves the efficiency of Monte-Carlo Localization System by using likelihood field, and can be applied to soccer robot as well.