Abstract:

The autonomous positioning of unmanned carrier-based aircraft on the aircraft carrier is an important prerequisite for realizing autonomous transfer and improving the efficiency of entry/exit. Among them， how to rely on airborne sensing equipment to realize autonomous positioning in the ship’s surface environment of carrier-based aircraft in GPS-denied environment is the key technology that needs to be solved urgently. Therefore， an autonomous positioning algorithm is proposed for unmanned carrier-based aircraft based on visual and lider fusion. The algorithm performs multi-sensor online calibration by combining the hand-eye calibration and mutual information calibration methods， so that the unmanned carrier-based aircraft can still run stably after the fuselage shakes in the face of wind and waves and the operation of the aircraft carrier causes the sensor external parameter calibration results to change. A factor graph is introduced for multi-sensor pose joint optimization of unmanned carrier-borne aircraft， and a simple and efficient sensor failure criterion is established based on the motion model in the process of autonomous transport， so as to integrate the lidar and visual positioning results effectively so that the algorithm can operate stably even when a single sensor fails. Finally， a multi-sensor-based simulation system of carrier-based aircraft on the ship is established to verify the algorithm. Experimental results show that the positioning error of the algorithm remains within 0.2 m when a single sensor fails， which meets the requirements of practical applications.

Key words: unmanned carrier-based aircraft, multi-sensor fusion, autonomous positioning, online calibration, simulation of the ship’s surface