Research On Visual/Inertial Based Adaptive Integrated Navigation

GPS(Global Positioning System)is currently the most widely used navigation system,but the stability of the GPS based navigation method relies on the stable reception of satellite signals,especially when satellite signals are interfered or blocked.Recently,Visual/inertial integrated navigation is widely used in daily life due to its adaptability.This paper mainly studies the visual/inertial based adaptive integrated navigation and the main work are summarized as follows:(1)The visual/inertial integrated navigation system is described and summarized.The research status at home and abroad is summarized and analyzed.The key problems of integrated navigation system are further pointed out.(2)An adaptive algorithm based on the invariant extended Kalman filter is proposed to address the inconsistency and the non-Gaussian noise in traditional integrated navigation approach.The main innovation lies on the modeling of the prediction and observation error based on Lie group,and adaptive invariant extended Kalman filter is proposed to estimate the position and state.During the filtering process,a fading factor based on innovation is calculated to dynamically modify the filtering gain,which finally improves the robustness of the system in a non-Gaussian noise environment.Furthermore,the effectiveness and superiority of the invariant extended Kalman filter based adaptive integrated navigation algorithm are proved by simulation experiments.(3)An adaptive unscented Kalman filter based on Lie group is proposed to improve the accuracy and stability of the visual/inertial integrated navigation system under the condition of data's abrupt change.The main innovation is that the position and state are modeled based on Lie group/Lie algebra,and the position and state are estimated by the proposed adaptive unscented Kalman filter.During the filtering process,the adaptive factor is calculated based on the innovation of Lie Group's UKF filter to dynamically modify the filter gain,which improves the robustness of the integrated navigation when data change abruptly.Moreover,in order to ensure the numerical stability in the filtering process,the square-root filter based implementation of the proposed adaptive algorithm is given.The effectiveness and superiority of the proposed algorithm is proved by simulation experiments.