Advancing Estimation Accuracy of Sideslip Angle by Fusing Vehicle Kinematics and Dynamics Information With Fuzzy Logic

In this paper, a vehicle sideslip angle (VSA) estimation method is presented and experimentally verified by fusing vehicle kinematics and dynamics information with fuzzy logic. First, a vehicle-kinematics-based (VK-based) reduced inertial navigation system (R-INS) is developed to calculate the VSA, velocities, and attitude. Then, with the velocities measured by a single-antenna global navigation satellite system (GNSS), a velocity-based Kalman filter (VBKF) is employed to estimate the velocity, attitude, and gyro bias errors in the R-INS and these errors will be adopted to correct accumulated errors of the R-INS. However, the heading error in the R-INS which is highly correlated with the VSA is not well observable when the vehicle is with low excitation. To address this observability issue, the vehicle-dynamics-based (VD-based) VSA estimation approach is leveraged to augment the heading error into the velocity measurements of the VBKF by a novel heading error measurement model, and an augmented Kalman filter (AKF) is designed. Next, according to the vehicle lateral excitation, a fuzzy logic method is proposed to fuse the heading errors from both the VBKF and AKF to take advantage of the VK-based and VD-based methods. Finally, a comprehensive experimental test is conducted, and the results confirm that the heading error observability issue in the VBKF has been tackled by fusing vehicle kinematics and dynamics and the VSA estimation accuracy has been advanced. The VSA estimation accuracy of the proposed method (the absolute mean error is only 0.126°) matches our previous work, which needs a dual-antenna GNSS, but the cost is reduced.