Low-cost AHRS design based on extending Kalman filter

A low-cost,high-accuracy attitude and heading reference system(AHRS)is proposed for the problem of high cost,big volume and high power consumption on the traditional AHRS which is applied in micro unmanned aerial vehicle(MUAV)and robots.Using a digital signal processor(DSP)as a hardware platform,9-degree of freedom micro-electro-mechanical system(MEMS)sensors including gyroscope,accelerator and compass are integrated.An attitude estimation method based on quaternion is adopted.Sensor model and state space model are built.The impact of acceleration on the system accuracy is considered.The singularity problem of quaternion covariance is solved,and the accurate output of attitude and heading by applying an extending Kalman filter to data fusion is acquired.The result indicates that the static accuracy and dynamic accuracy of the attitude angle are better than 0.5°and 2°,respectively.The system is also tested on the MUAV and the outcome meets the application demand of MUAV.