A Structurally Constrained Informer for UWB Positioning Distance Filtering Used in Speed-Skating Training

Scientific training for location and trajectory tracking technology is a necessary technical condition in speed skating project, this paper studies the crucial data filtering technique for achieving a good trade-off between system cost and the positioning accuracy based on UWB positioning method. This paper focuses on the problem of localization signal blocking caused by metal reflection in large stadiums, and an improved trajectory filtering method based on Transformer model has been proposed with adjusted novel structural constraints. Compared with the traditional Kalman filtering method, our proposed method achieves prediction accuracy improvement up to 68%, and compared with the original Informer model, the prediction accuracy is improved by 19%. It makes it possible to track accurate positions and generate trajectory information in crowded and intense race conditions.