Enhancing Mobile Robot Location Estimates Using Sensor Fusion

This paper investigates the use of multiple sensor types in the generation of optimal estimates of the position and orientation of a mobile robot within a structured environment. A novel application of the extended Kalman filter algorithm is used to combine observations of the robot’s surrounding environment from various sensor types with the measured wheel rotation of the robot to produce accurate estimates of a mobile robot’s position and orientation. A simulation of a mobile robot navigating along a complex trajectory is presented where the robot’s position and orientation estimates are produced via vision-based as well as range-based observations of the robot’s surrounding environment. The results of the simulation show that estimation accuracy can be increased through the use of multiple sensor types when compared to a system which uses only a single sensor type.