Multi-strategy hybrid particle swarm algorithm for magnetometer error calibration.

Abstract

To address the accuracy degradation caused by inherent errors in fluxgate magnetometers, this study proposes a Multi-Strategy Hybrid Particle Swarm Optimization (MSPSO) algorithm. This method effectively balances global search scope with local search depth, overcoming the limitation of conventional Particle Swarm Optimization (PSO) algorithms that tend to become trapped in local optima, and achieves high-precision, highly robust magnetometer calibration. Experimental results demonstrate that compared to PSO, modified particle swarm optimization, dynamic hierarchical elite-guided particle swarm optimization, and robust ellipsoid fitting methods, MSPSO reduces the average root mean square error by 73%, 54%, 41%, and 49%, respectively. This work provides a reliable solution for magnetometer calibration.