Research On Online Estimation Method Of Inertial/satellite Integrated Navigation Parameters

The accuracy of inertial devices and satellite signal receiver must be described when the two information are fused. These inaccuracies are generally treated as noise, which is unknown. Consequently, how to properly set the noise covariance is heavily depend on experience in the engineering application. This will lead to lower filter performance and even filter divergence, because the noise may change in flight condition. Moreover, the integrated navigation is not accuracy enough in error angles and other states variables. It caused the velocity error and position error, which is amended, going far away in a short period. So, to improve performance of integrated navigation, this paper explored what is the main factor limiting integrated navigation accuracy and how to setting related parameters online.Thus this paper mainly studies as follows, strapdown inertial navigation equations are established firstly. Then analyzing the law of error propagation and deriving filter model. A new hardware structure is presented as an enhanced integrated navigation part. Based on the new structure, the observation of state variables is analyzed. The fatal factors constraining navigation accuracy were found out by using formula deriving and matlab emulation. Genetic algorithm is introduced to estimate the parameters. To play the advantages of GA, uniform design is used to generate initial population instead of random numbers, and a fitness function is studied to judge which the optimal individual is and when to stop estimating. At last, the validity of the proposed approach is confirmed by simulations.