基于边界归一化的低空无人机实时避撞路径规划

摘要

小型无人机的广泛应用已引发严重的城市低空安全问题,实时避撞路径规划是提高无人机在陌生城市低空飞行安全的重要技术.针对传统Laguerre路径规划无法实时应用的问题,建立了基于单航段分解的Laguerre实时滚动规划模型.通过障碍边界归一化确保障碍物异型位置关系下的Laguerre路径的一致性,导出了单航段Laguerre最佳避撞路径规划定理.进而,提出了基于障碍边界归一化的无人机低空实时避撞路径规划方法.通过模拟城市复杂环境的规划实验表明,该方法不仅实现了陌生环境下的实时路径规划,而且所规划路径具有更好的避撞安全性.%Due to widespread application of small UAVs, serious safe problems for urban lower aerial areas have being risen.For an unknown urban area, the real-time collision avoidance route plan is the important technology for improving the UAV lower aerial flight safety.Aiming at the problem that the traditional Laguerre route planning cannot be used for real time application, a real-time rolling model for Laguerre planning based on decomposed single flight segment is firstly built in this paper.Furthermore, by obstacle boundary normalization, the consistency of Laguerre route is ensured for various position relations among obstacles, and the safest Laguerre route planning theorem for collision avoidance for single flight segment is exposed.Based on above analysis, a real-time route plan method for UAV lower aerial collision avoidance based on obstacle boundary normalization is presented.The planning tests have been done for simulating complex urban district, the testing results show that the presented route planning method can not only be applied for real-time route planning in unknown areas, but its collision avoidance safety for planned route is better than other compared methods.