Self-Position Tracking Based on Bidirectional Filtering Under Beacon and Angle Measurement Perturbation
TL;DR
This article derives an approximate closed-form solution based on total least squares (TLS), which can suppress both beacon and angle measurement errors and can achieve better performance than other methods, including the clustering algorithm with weighted moving average (WMA) and the pseudolinear estimator algorithm with WMA.
Abstract
The Global Navigation Satellite System (GNSS) signal will be weakened in extreme environments, which may cause the decreased GNSS precision. In this article, we propose a self-position tracking method under noncooperative beacon and angle measurement perturbation. First, for eliminating abnormal angle measurements, we propose an outlier data removal method based on farthest distribution angle search and singular value reconstruction. Then, we derive an approximate closed-form solution based on total least squares (TLS), which can suppress both beacon and angle measurement errors. In addition, for directional deviation problem in tracking, the multilayer yaw orthogonal projection is developed to perform multilayer projections of TLS positioning results onto yaw angles, which can reduce error dimensions. The moving average with prior information and the Kalman filter without state parameters are constructed to relieve distance errors in yaw dimension, and the bidirectional filter utilizing the mean of forward and backward filtering is proposed to smooth the whole trajectory. Both simulated and practical experiments show that the proposed method can achieve better performance than other methods, including the clustering algorithm with weighted moving average (WMA) and the pseudolinear estimator algorithm with WMA.