Uncertainty evaluation and optimization of INS installation measurement using Monte Carlo Method

Abstract

Purpose – The purpose of this paper is to increase the measurement accuracy of assembly deviations of an inertial navigation system, a new evaluation and optimal method of assembly metrology system is proposed, which takes into account the uncertainty from laser tracker hardware and coordinate system transformation, and is based on the Monte Carlo method. Design/methodology/approach – The uncertainty model of the laser tracker is established and its parameters are obtained from the known repeated test data by kriging interpolation and the least squares method. The errors of coordinate transformation are reduced by using a weighted point matching method, and the uncertainty of the transformation parameters is obtained based on the generalized inverse theory. The weighting coefficients of each reference point are optimized by the particle swarm optimization method according to the assembly requirements. Findings – The experiment results show that measurement error and predicted results match well, and the assembly deviation uncertainty of large component is reduced by about 10 per cent compared with the singular value decomposition method. Originality/value – This paper proposes a method to evaluate and eliminate the influence of random errors of the laser tracker during evaluation process of coordinate translation parameters and assembly deviations. The proposed method would be useful to improve the assembly measurement accuracy through less measurement times.