A Hybrid Method for Positioning a Moving Magnetic Target and Estimating Its Magnetic Moment
Cheng Chi, Dan Wang, Zhentao Yu, Fajie Wang, Lu Yu, Feng Qin
Abstract
To solve the problem of false solutions being present in the inversion of magnetic target parameters on airborne scalar magnetometer and of the inversion method being limited to stationary targets, this article proposes a hybrid method for estimating both the motion parameters and magnetic moments of single moving magnetic target. The proposed method first measures the magnetic field signals along a cross-crossing flight trajectory and then denoises the measurement signals using a signal extraction algorithm based on orthonormal basis. After that, it builds an optimization model based on magnetic dipole model and the geometric relations between the magnetic target and the magnetometer. To solve the problem of the model easily converging to locally optimal solutions, a hybrid algorithm combining the adaptive weighted particle swarm optimization (AWPSO) algorithm and the Levenberg–Marquardt (LM) algorithm is proposed to determine the target parameters. The hybrid algorithm first uses the AWPSO algorithm to produce initial values for the target parameters via iteration, and then, the initial values of the parameters are used as the input to determine the precise values of the target parameters by fitting the LM algorithm. Simulation and actual test results show that the proposed method has high positioning accuracy and good robustness. Simulation shows that when Gaussian white noise with a standard deviation of 0.2 nT is superimposed, the max relative error of the target parameter estimation is 4.6%. The field test results of a water surface ship show that the average positioning error of the method is 20.9 m.