Simulation of mobile robot motion considering uncertainties in robot model

This paper describes a simulator which emulates mobile robot motion. The simulator implements probabilistic model of robot motion and sensor output. The sensor implemented is a range sensor such as laser range finder which is used widely for obstacle detection, map making, and localization. The sensor model includes measurement noise with Gaussian distribution around true range value, failure of detection, and unexplainable random measurement. The motion model incorporates uncertainty with normal distribution. The simulator is used to test a collision avoidance algorithm. In practical application of the mobile robot, major problems arise due to the unexpected uncertainty of motion and sensor. The simulator is useful to predict the problems of robot motion before the robot is deployed in the field. For wider availability, the simulator can be further developed to include error model of various sensors.