Information Monte Carlo localization algorithm for fusing distributed perception

A distributed perception network is designed including environmental cameras and a laser sensor on the robot, and accordingly, an information Monte Carlo localization algorithm is proposed to fuse distributed perception. In the process of Monte Carlo localization, an efficient numerical method is employed to approximating the information utility in sensor selection. Consequently, the robot pose is updated according to the measurement from the optimal sensor node. Experimental results illustrate the validity of the method in solving problems of global localization and “kidnapped robot”.