Abstract: High-precision magnetic positioning has been widely used in surgical navigation, capsule endoscopy, and other medical engineering applications, and has become an important topic in the Internet of Things. According to the magnetic source, localization systems can be classified into alternating-current coil-based and permanent-magnet-based schemes. Owing to their compact structure, permanent-magnet-based systems are more suitable for portable and embedded applications. Compared with single-rotating permanent magnets, biaxial rotating permanent magnets provide richer magnetic field information and greater potential for six-degree-of-freedom pose tracking. However, accurate estimation in such systems remains difficult because of the nonlinear magnetic field model, symmetry-induced ambiguity, initialization sensitivity, and time-varying measurement uncertainty. To address these issues, this paper develops a measurement-oriented framework for biaxial rotating permanent-magnet pose tracking, including an observability-enhanced excitation scheme, a physically consistent initialization strategy, and an adaptive recursive reconstruction method based on trust-region optimization and residual-corrected unscented Kalman filtering. The proposed method is validated through Monte Carlo simulations, COMSOL-based finite-element analysis, and prototype experiments. The results demonstrate improved convergence reliability, reduced sensitivity to initialization and environmental uncertainty, and better repeatability and robustness in complex tracking scenarios.

Similar articles