Research On SINS Experiment Design Method Based On Multi-objective Optimization

Strapdown Inertial Navigation System(SINS)is currently a very common navigation system,and its navigation accuracy is affected by many factors.Among them,the accuracy of inertial sensitive components(accelerometer and gyroscope)is the main factor affecting the accuracy of SINS.In order to improve the accuracy of inertial sensitive components,an error model is usually established and the error coefficient is calibrated.The accuracy of the identification of the error coefficient greatly affects the accuracy of inertial sensitive components.The error coefficients of the error model include static error coefficients and dynamic error coefficients.However,compared with static error coefficients,dynamic error coefficients are more difficult to accurately calibrate and their impact on navigation accuracy cannot be ignored.Therefore,designing a good experimental plan can improve the accuracy of the error coefficient and thus the SINS accuracy.This paper proposes an optimal test design method based on the improved archived multi-objective simulated annealing algorithm(IAMOSA)for the dynamic error coefficient calibration test of SINS inertial sensitive components on a dual-axis test turntable.In this paper,the error model of the accelerometer and gyroscope of SINS under the calibration test of the dual-axis turntable is established,and a continuous linear time-varying system model containing 9 accelerometer dynamic error coefficients and 9gyroscope dynamic error coefficients is established according to the dynamic equation.After discretizing the continuous system model,the error coefficient estimation for the experiment using discrete Kalman filtering is determined according to the optimal estimation theory of parameter calibration.The estimated performance of Kalman filtering is analyzed,and two optimization goals for experimental design are proposed in terms of improving the estimation accuracy of error coefficients and improving the efficiency of experiments.The optimization problem of SINS experimental design is transformed into a multi-objective optimization problem.This paper also studies and analyzes the multi-objective optimization algorithm,and the existing archived multi-objective simulated annealing algorithm(AMOSA)is improved to obtain the improved algorithm IAMOSA to improve the convergence and distribution of the algorithm.First,the algorithm is improved by adding a large new archive set.The new archive set saves the non-dominated solutions deleted by theclustering algorithm.At the end of the algorithm,it is merged with the solution set obtained by the algorithm and then the fast non-dominated sorting algorithm is used to obtain the final solution set.This can improve the distribution of the solution.Second,by adding a pole set to identify and protect the poles,the clustering algorithm of AMOSA is improved.This reduces the probability of the frontier loss of the algorithm,improves the coverage of the frontier,and thus improves the convergence and distribution of the algorithm.At the end of this paper,IAMOSA algorithm is used to solve the constrained experimental design optimization problem,and its effectiveness is verified.The estimated errors of the dynamic error coefficients of the inertial sensitive components are analyzed,and it is verified that the IAMOSA algorithm can not only effectively improve the identification accuracy of the error coefficient,but also provide decision makers with a variety of experimental plans with different identification accuracy.