Quaternion attitude and disturbance observer mass estimation with low-cost sensors for industrial vehicles with trailers

Abstract

Information about the vehicle mass and ground slope is important for many assistance functions in road and logistics vehicles. In intralogistics, safe speed limits for tractors depend on both slope and trailer mass. This work presents an estimation procedure for the attitude of a vehicle and the mass of a trailer attached to it. The estimator works in two steps. First, the attitude is estimated using an extended Kalman filter based on acceleration and angular rate measurements complemented by a single track vehicle model. In the second step, this information is used together with velocity and motor torque data to estimate the trailer mass and friction coefficient. A Kalman filter based disturbance observer for parameter estimation is compared to a recursive least squares identification. The mass estimation is extended by a structural break test to detect sudden mass changes when hitching and unhitching trailers. Multiple variants of the estimation scheme are implemented on an intralogistics vehicle. The performance of the proposed attitude and mass estimation solutions is demonstrated in comparison to state of the art reference algorithms in a large number of experiments. Compared to state of the art estimators, the proposed estimator yields a median 54 % reduction of pitch estimation RMSE and 40 % reduction of the mass estimation error. The structural break detection is able to detect all instances of hitching and unhitching of trailers with few false positives.