Attitude and heading information fusion method based on linear Kalman filtering

Abstract

The invention provides an attitude and heading information fusion method based on linear Kalman filtering. The method is characterized by comprising the steps that a low-precision inertial navigationerror equation is established, and a relation between misalignment angles and horizontal accelerations and a linear Kalman filter model are established; next, gyro measured values are used to performattitude update, acceleration measured values are converted into navigation coefficients, an acceleration error is calculated, and estimated misalignment angles are used to correct an attitude quaternion to realize real-time estimation-correction; and last, the quaternion is converted into an Euler angle for display. Through the method, the problems that through an EKF algorithm, the calculation quantity is large, and it is difficult to analyze and solve a state vector partial derivative are solved; meanwhile, a proportionality coefficient can be adjusted in real time, and gyro zero offset canbe estimated; a first-order Markov model established for a gyro avoids the problem of excessive convergence of a filter; and the method is superior to a complementary filtering algorithm and a gradient descent algorithm in performance, is equal to the EKF algorithm in performance and inferior to the EKF algorithm in calculation quantity and programming complexity and is relatively suitable for engineering realization.

Description

technical field [0001] The invention relates to the field of robot SLAM, in particular to a heading information fusion method based on linear Kalman filtering. Background technique [0002] SLAM (simultaneous localization and mapping, real-time positioning and mapping) is the key to realize fully autonomous mobile robots. Among them, the core of localization (positioning) is to know the kinematic parameters such as the robot's heading, attitude (hereinafter referred to as heading), speed, and position. The current attitude generally uses IMU (InertialMeasurement Unit, Inertial Measurement Unit), and cooperates with the previous attitude algorithm to calculate the attitude information of the robot. [0003] The basic idea of ​​the attitude algorithm is to fuse the approximate horizontal attitude angle information obtained by the accelerometer information of the IMU and the attitude information obtained by the gyroscope integration. The reason for fusion is that the attitude...

AI Technical Summary

Problems solved by technology

This method solves the problem that the EKF algorithm has a large amount of calculation and the partial derivative of the state vector is difficult to obtain analytically. At the same time, it has the characteristics of real-time adjustment of the proportional coefficient and estimation of the zero bias of the gyroscope; and a first-order Markov model is established for the gyroscope. , to avoid the problem of filter transition convergence; its performance is better than that of complementary filtering and gradient descent algorithm, and its performance is equivalent to that of EKF algorithm, but its calculation amount and complexity are smaller than that of EKF algorithm. method

Images