无人机群体视角下的轨迹预测

Trajectory prediction is a key task in unmanned systems. Current research on the trajectory prediction is based on the trajectory dataset collected by a single device, commonly by UAV. Due to the influence of obstacles and occlusion in the scene, it is difficult to determine the specific location of the target in complex scene and tends to lose the target. In the meantime, current works in trajectory prediction observed by UAV usually rely on a bird's-eye view (BEV), which restricts the flexibility of UAV. The development of group collaboration technology brings new opportunities to overcome these limitations of single UAV. From the group perspective, UAVs can effectively solve the problems of target occlusion and target localization. In the meantime, the pose estimation based on UAV group can compute the accurate 3D location of targets and provide a basis for feasible observation of the scene. In this paper, starting from the related work of trajectory prediction from the group perspective, we aim to present the new topic of trajectory prediction from UAV group perspective and discuss the challenges and solutions of this field, and provide some potential insights for the trajectory prediction research.