Abstract:

The width of the virtual tube directly affects the flow rate of unmanned aerial vehicles in the tube. However, current virtual tube planning methods cannot guarantee the tube width, which leads to blockages that severely impact the efficiency of unmanned aerial vehicle passage. In this regard, a method for planning a virtual tube space is proposed that can generate a tube generation path meeting the virtual tube space requirements in an environment containing obstacles. The method mainly includes path searching and trajectory optimization. It is based on the rapidly exploring random tree* (RRT*) search, where space detection is used to determine whether there is sufficient space around each segment of the path. The search is then re-expanded, replacing path segments with insufficient space if the space is not enough. In the trajectory optimization process, a uniform B-spline is used to parameterize the trajectory, and a collision cost function and a smoothing cost function are designed to keep the trajectory away from obstacles, providing enough space for virtual tube planning. Simulation tests verify the superiority and robustness of the proposed method in large-scale swarms, with the reduction of 20% of average passage time of a swarm containing 15 unmanned aerial vehicles.

Key words: virtual tube, quadrotor aircraft, motion planning, path searching, uniform B-spline, space satisfaction