Maneuvering Acceleration Assisted Attitude Algorithm Design Based on Fuzzy Adaptive Kalman Filter

Problems that accelerometers have wrong correction on horizontal attitude of maneuvering aircraft and that noise statistical properties change with the actual working conditions in low accuracy Attitude and Heading Reference System (AHRS), are investigated in this paper. Firstly, the maneuvering acceleration is described with current statistical model. Secondly, a 9-state EKF is introduced with 9 states of three attitude angles error, triaxial gyro bias error and triaxial maneuvering acceleration error. The observations include triaxial acceleration error. Finally, the proposed algorithm is designed based on fuzzy adaptive Kalman filter, which estimates and modifies parameters of observation model. Simulation results demonstrate that the data fusion and filtering methods are valid for horizontal attitude optimal estimation in low accuracy AHRS.