Abstract: The estimation of road slope is of great significance for the accurate control of vehicles and the construction and analysis of driving environment. In the previous slope estimation, the slope change rate is not considered, so it is difficult for the estimated value to follow the change of road slope in real time. And, the estimation results based on a single estimation method will be affected by braking, shifting, rapid acceleration and other vehicle states, so it is difficult to ensure the reliability and accuracy of slope estimation. Therefore, a road slope estimation method based on dual Kalman filter and probabilistic nearest neighbor data association filter is proposed. According to different vehicle states, the two sub estimated slope values based on dynamics and kinematics are fused globally to avoid the error of single sub-estimated slope value. And, for each sub estimation method, a slope change rate and slope hierarchical estimation algorithm based on dual (unscented) Kalman filter is proposed. Simulation and experimental results show that the proposed sub-estimation method can better follow the change of road slope and improve the estimation accuracy. The global fusion method can further improve the estimation accuracy on the basis of the sub-estimation method, avoid the large error of a sub estimation, and improve the accuracy and reliability of slope estimation.

Key words: slope estimation, dual Kalman filter, probabilistic nearest neighbor data association, estimation fusion