A hierarchical estimation algorithm for heavy-duty vehicle mass and road grade based on UKF and RLS

Real-time estimation of vehicle mass and road grade is essential for intelligent vehicle control. This study proposes a hierarchical sequential estimation method tailored for heavy-duty vehicles. In the first layer, vehicle mass is estimated shortly after the vehicle starts, and in the second layer, road grade is estimated based on the previously determined mass. To address challenges during non-normal conditions, such as braking and shifting, the method employs recursive least squares (RLS) and the unscented Kalman filter (UKF) under normal conditions. During non-normal conditions, the vehicle mass estimation retains the value determined prior to the event, while road grade is predicted using an ARIMA model based on historical grade data. Real-vehicle experiments show that the vehicle mass estimation error is less than 4.2%, and the road grade estimation achieves an RMSE of less than 0.2°, demonstrating a significant improvement in accuracy.