Robot autonomous navigation based on multi-sensor global calibrated

Aiming at the robot problems such as little relevance, low accuracy of the simultaneous localization and map building (SLAM) and easy locking, lack of initiative of the navigation system, the multi-sensor vision system is introduced, and then unifying the data of each sensor by world coordinate system of global calibration based on the local calibration of each vision sensor module, a serial of local maps are combined into a global map by the derive of Least-Square (LS). Preprocessing of the global map data is done by the genetic programming (GP) arithmetic and inference is done with the delta fuzzy rule to plan the best routine to achieve robotic autonomous navigation. Simulation results show that the robot can create accurate and complete map of the environment and bypass the obstacles agilely to reach the destination smoothly and reliably with the map. Thus the feasibility and effectiveness of this strategy is verified.