Abstract:
In future lunar surface roaming and inspection missions, considering the some adverse factors on the lunar surface, such as rugged road conditions, lack of structured scenes, poor surface texture and so on, a method of active re-observation of historical landmarks was proposed to improve positioning accuracy of lunar rover lidar. Firstly, the significance of the point cloud landform in the current detection field was studied and judged at the fixed time, and re-observed landmarks were extracted. The trigger time of re-observation was determined according to the position and attitude estimation of the real-time monitoring inspector. Finally, point cloud matching algorithm was used to obtain the accurate position and attitude. Simulation results show that the active re-observation method fuses the historical position and attitude estimation of the system with uses the historical position and attitude estimation of the system with current observations, suppresses data drift caused by noise, and improves positioning accuracy of the lunar rover.