Parameter Adjustment Method of Kalman Filter Based on Quadratic Curve Fitting

This paper proposes a method for adjusting parameters of Kalman filter based on quadratic curve fitting. Among them, the two main processes of the Kalman filter parameter adjustment are to obtain the true value of the target and how to fit the observation noise in the Kalman filter based on the quadratic curve. First of all, this paper realizes the installation and calibration of the sensor based on the installation characteristics of the sensor. Secondly, it realizes the transformation between the sensor coordinate system and the reference coordinate system based on the stereo calibration principle in computer vision, and uses Real-Time-Kinematic (RTK) to obtain the real position of the target in the scene. Then the observation noise in the Kalman filter is fitted based on the quadratic curve, and the Kalman filter parameters are adjusted based on the observation noise. Finally, the effectiveness of the method proposed in this paper is verified based on qualitative and quantitative methods. Experimental results show that the method proposed in this paper can effectively realize real-time Kalman filter parameter adjustment.