Garrido Santiago, Muñoz Javier, López Blanca, Quevedo Fernando, Monje Concepción A, Moreno Luis
Robotics Lab, Department of Systems Engineering and Automation, Universidad Carlos III de Madrid, Av. Universidad 30, 28911 Madrid, Spain.
Sensors (Basel). 2022 Apr 21;22(9):3174. doi: 10.3390/s22093174.
This paper studies the Fast Marching Square (FM2) method as a competitive path planner for UAV applications. The approach fulfills trajectory curvature constraints together with a significantly reduced computation time, which makes it overperform with respect to other planning methods of the literature based on optimization. A comparative analysis is presented to demonstrate how the FM2 approach can easily adapt its performance thanks to the introduction of two parameters, saturation α and exponent β, that allow a flexible configuration of the paths in terms of curvature restrictions, among others. The main contributions of the method are twofold: first, a feasible path is directly obtained without the need of a later optimization process to accomplish curvature restrictions; second, the computation speed is significantly increased, up to 220 times faster than other optimization-based methods such as, for instance, Dubins, Euler-Mumford Elastica and Reeds-Shepp. Simulation results are given to demonstrate the superiority of the method when used for UAV applications in comparison with the three previously mentioned methods.
本文研究快速行进正方形（FM2）方法，将其作为无人机应用中的一种具有竞争力的路径规划器。该方法满足轨迹曲率约束，同时显著减少计算时间，这使其在基于优化的文献中的其他规划方法方面表现更优。进行了对比分析，以展示FM2方法如何通过引入两个参数（饱和度α和指数β）轻松调整其性能，这两个参数允许在曲率限制等方面灵活配置路径。该方法的主要贡献有两方面：第一，无需后续优化过程即可直接获得可行路径以满足曲率限制；第二，计算速度显著提高，比其他基于优化的方法（如杜宾斯方法、欧拉 - 芒福德弹性曲线法和里兹 - 谢泼德方法）快达220倍。给出了仿真结果，以证明该方法用于无人机应用时相对于上述三种方法的优越性。
