An observation algorithm for key motion states of skid-steered wheeled unmanned vehicle

With the increasing demand of military and civilian in the intelligent vehicles, the skid-steering theory has been widely used in unmanned ground vehicles, especially in unmanned military vehicles and unmanned surveillance platforms. Due to its driving environment complex and variable, which requires stricter dynamic control system. In order to improve the active safety performance of the skid-steering unmanned vehicle and develop the key technologies such as behavior decision planning technology, path tracking, and dynamic control technology, it is necessary to develop the dynamic state parameter observation system based on skid-steering theory. In this paper, an observation using Strong Track External Kalman Filter theory with noise matrix adaptive is designed to estimate vehicle kinematic parameters based on a 6 × 6 skid-steered unmanned vehicle. First, kinematic and dynamic model is built to analyze the characters of a skid-steered wheeled vehicle. Then a tire force estimation method based on dynamic model is presented to observe the tire longitude and vertical force. The tire force data is also used by Dugoff nonlinear model. Then an External Kalman Filter theory is designed to estimate vehicle kinematic parameters. To increase the accuracy and the robustness of the observer, the Strong Tracking EKF (STEKF) and noise adaptive adjustment is designed. Finally, a combined simulation using TruckSim and Simulink and the experiment using a 6 × 6 skid-steered unmanned vehicle verifies the efficiency of the observer. Results show that the observer is able to estimate the skid-steered wheeled vehicle states, and it also shows that the yaw rate result in the slip angle difference between each tire.