Map building of unknown environment based on fuzzy sensor fusion of ultrasonic ranging data

This paper investigates the use of ranging data collected from ultrasonic sensors mounted on a two-wheeled mobile robot, Pioneer 3-DX, to build an occupancy grid map for an unknown indoor environment based on fuzzy sensor fusion of the ultrasonic ranging data. Because of uncertainties inevitably encountered by using the ultrasonic sensors, a more reliable sensor model is derived to solve the problems of angle uncertainties and multiple reflections. To address the problems due to measurement uncertainties of the ultrasonic sensors, a fuzzy logic approach is proposed to construct the grid map, where the information of the grids is continually computed and updated through fuzzy logic operations. As long as the environmental map in which every grid is represented as the fuzzy degree of occupancy is obtained, it can be used for localization or path planning to enhance the navigation autonomy of mobile robots. To validate the feasibility of the proposed approach, several experiments are conducted to build maps for various indoor environments.