Improved BP-Based Cooperative Localization Algorithm of Factor Graph in UAV Networks

In this paper, an improved cooperative localization (CL) method based on factor graph for Unmanned Aerial Vehicle (UAV) networks is proposed. Belief propagation (BP) is employed to realize CL and UAV nodes exchange their information on the given factor graph model. We improve the traditional non-parametric belief propagation (NBP) algorithm by applying different optimal algorithms such as particle filter (PF) or unscented kalman filter (UKF) in different iterations of BP in order to reduce the computational complexity. In the simulation, performances of the extended kalman filter (EKF), UKF, traditional NBP, and the proposed method are compared. Simulation results show that the improved algorithm has the lower root mean squared error (RMSE) than the other algorithms and lower computational complexity compared with the traditional NBP algorithm.