基于因子图优化和李群空间表征的室内行人协同导航

Providing navigation and localization services for pedestrians in indoor environments has always been challenging. To address the problem of pedestrian formation navigation in indoor environments without infrastructure, a collaborative indoor pedestrian navigation based on factor graph optimization and Lie group spatial representation is proposed. The two-dimensional special Euclidean-group (SE(2)) space in Lie group is used to describe the motion state of pedestrians, and multi-user and multi-calendar PDR/UWB observation information is used to optimize the optimal estimation of pedestrian formation motion state by nonlinear solution. The experimental results show that compared with the PDR and EKF algorithms, the average positioning error of the proposed algorithm is reduced by 34.4% and 10.4%, respectively, and the relative distance error of the formation is reduced by 76.2% and 45.4%, respectively, which can effectively improve the positioning accuracy of the pedestrian formation.