基于混合动态信念传播的多无人机协同定位算法

针对多无人机(UAVs)协同定位问题,提出一种基于混合动态信念传播的定位算法。在部分无人机GPS信号丢失的情况下,该算法可根据其他无人机的GPS观测,相邻无人机之间的相对距离观测,以及无人机加速度计的输出,对每个无人机的位置和速度状态进行分布式在线估计。首先用因子图模型描述多无人机的联合信念状态,接着给出一种混合动态信念传播推理算法计算图模型中的每个变量节点(对应于每个无人机)状态的边缘后验分布。推理过程仅包括每个无人机对自身局部信息的处理以及相邻无人机之间的信息交互,因此该算法可完全分布式实现。通过仿真实验以及与传统协同定位算法的比较,表明了本文算法的有效性。

Cooperative localization algorithm of multi-UAVs based on dynamic hybrid belief propagation

A dynamic hybrid belief propagation algorithm for cooperative localization of multiple unmanned aerial vehicles (UAVs) was proposed. In the case that the GPS signals of some UAVs are missing, the position and velocity of each UAV in the group in an online and distributed manner based on the GPS observations of other UAVs, the relative distance measurements between neighboring UAVs, and the outputs of accelerometer of each UAV were estimated by the algorithm. Specifically, the joint belief state of multiple UAVs by factor graph was modeled, and the marginal posterior distribution of each node in the graph was calculated, corresponding to each UAV, using the hybrid dynamic belief propagation algorithm. The inference only involves local information processing on each UAV and mutual information exchanging among neighboring UAVs, allowing the algorithm to be implemented in a completed and distributed manner. The results show that the effectiveness of the method by comparison with traditional cooperative localization algorithms in simulations.