A hierarchical collaborative dynamic particle swarm optimization method for triaxial magnetometer error calibration

Purpose This study aims to improve the calibration accuracy of triaxial magnetometers by developing a hierarchical collaborative dynamic particle swarm optimization (HCD-PSO) algorithm for error parameter estimation. Design/methodology/approach An HCD-PSO algorithm is proposed for triaxial magnetometer error parameter estimation and calibration. The proposed method improves the standard particle swarm optimization (PSO) framework by incorporating an adaptive parameter collaborative regulation mechanism, a low-quality particle information source reconstruction mechanism, and a dual-population competitive collaborative evolution mechanism, thereby enhancing population diversity and improving convergence performance and parameter estimation accuracy. Findings Simulation results demonstrate that HCD-PSO can achieve stable error estimation results with low calibration errors under different noise conditions. Comparative experiments further show that HCD-PSO outperforms PSO, two-Swarm learning particle Swarm optimization and Sand Cat Swarm Optimization in terms of parameter estimation accuracy and convergence efficiency. In addition, parameter sensitivity analysis confirms that the proposed algorithm maintains stable optimization performance under different parameter search range settings. Additional comparisons with two classical calibration methods, further show that HCD-PSO achieves lower magnitude MAE and RMSE. Real-sensor experiments further show that HCD-PSO reduces the magnitude MAE from 11.347 µT before calibration to 0.116 µT after calibration. Originality/value The proposed method effectively improves the accuracy of triaxial magnetometer error parameter estimation and provides a reliable intelligent optimization approach for high-precision magnetometer calibration.