This work proposes a multi-kernel correntropy Kalman smoother for orientation estimation and considers position constraints for trajectory estimation using foot-mounted IMUs. We prove that the maximum positioning error can be reduced to 25% of the original values by involving a single loop closure, and it can be reduced to a designated value by designing multiple position constraints properly.
Simulation results of the proposed multi-kernel correntropy Kalman smoother are available in simulation/smoother_impulsive_measurement.m.
Experimental results of the square experiment are available in Square_walking/walking_square.m.
Experimental results of the stair walking are available in Stair_walking/stair_walking_rfoot_lfoot.m.