Abstract
This paper presents that the method of position error correction for planetary rovers by using skyline matching azimuth estimation. Skyline matching is the method that estimating position and attitude by comparing two skylines. One skyline is simulated one from a digital elevation model (DEM). The other one is gotten from camera images taking around view. We focus on using azimuth attitude estimated by skyline matching to correct accumulation error of IMU. We modeled IMU angular velocity bias increases linearity. We conducted field experiments in Oshima Island and Chiba. As a result, estimated relative azimuth errors are lower than 1 degree. IMU azimuth errors decreased by 67% (6.25 deg to 2.15 deg ) and IMU position errors decreased by 70% (3.33m to 1.04m).
