An adaptive aiding algorithm for pedestrian navigation

This paper presents a novel aiding algorithm for pedestrian navigation using foot-mounted inertial measurement units (IMUs). Autonomous pedestrian navigation with foot–mounted IMUs is based on the integration of simplified navigation equations and the correction of the navigational solution with zero velocity. Additional aiding algorithms are needed in the absence of external information such as GNSS or Wi-Fi and Bluetooth signals. There are two main groups of such algorithms: aiding based on information about bounded step length (two IMUs on both feet are required) and aiding based on straight-line path detection (heuristic drift elimination, HDE). The first method does not consider different accuracy of IMUs whereas the performance of the second one strongly depends on trajectory form. An attempt to eliminate the drawbacks of both algorithms is undertaken below. The novel algorithm is an adaptive version of the method based on bounded step length. Adaptivity is provided by tuning the measurement matrix for the less accurate IMU. The accuracy is assessed through the trajectory analysis based on information about straight-line motion. The novel algorithm is tested on experimental data. According to the testing results, this algorithm has better performance in the experiments with complicated trajectories. It can be used within an integrated pedestrian navigation system in the absence of external information.