Quadruped robot roll and pitch estimation using an unscented Kalman filter

Abstract

We present a novel algorithm for estimating a quadruped robot's pitch and roll angles. Assuming even terrain and an ideal bounding gait, we compute the roll and pitch angles to be fused with on-board Inertial Measurement Unit (IMU) measurements in an unscented Kalman filter (UKF). Simulation results illustrate the validity of the methodology developed. It is shown that the error in the estimation of both angles is much smaller compared to those in the literature.