Teleoperation of unmanned ground vehicles based on 3D trajectory prediction

We report an automated method for teleoperation of unmanned ground vehicles in complicated environments using multi-channel LiDAR and a video camera. To improve accurate remote control and compensate signal transmission latency, a method for trajectory prediction based on dynamic model is proposed. Unlike earlier 2D trajectory prediction, our method utilizes LiDAR point data to generate 3D trajectory. Our approach enables capabilities for teleoperation, including fused data displays and multiple remote control modes. This system has been evaluated through experiments on real world datasets of various environments.