An attitude determination system based on adaptive kalman filter for human machine interface

Abstract

This paper presents a low cost three degree-of-freedom attitude determination system for human machine interface. The system consists a triaxis accelerometer and a triaxis magnetometer. The attitude measurement algorithm which direct calculates the attitude is first discussed, and the error due to motion acceleration is analyzed, then an adaptive Kalman filter which dynamically modifies the measurement variance is proposed to estimate the attitude. The proposed filter is implemented on a homemade module, and compared with commercial motion tracking system which uses an additional triaxis gyroscope. Experimental results show that this method can be used for low-cost real time human machine interface. © 2011 Springer-Verlag.