Ranging Error Suppression Algorithm Based on CNN-Transformer-BiLSTM

This article proposes a distance measurement error suppression algorithm based on convolutional neural network (CNN)-Transformer-Bidirectional long short-term memory (BiLSTM), aiming to suppress distance measurement errors in indoor positioning, thereby improving indoor positioning accuracy. This article combines the feature extraction ability of CNN, the global dependency captures ability of Transformer, and the bidirectional sequence modeling ability of BiLSTM to construct an efficient ensemble learning framework. Through the verification of measured ultra-wide band (UWB) data and simulation data, the algorithm can significantly reduce ranging errors. The average ranging error is reduced by about 45.61% and 61.62% in the measured and simulation datasets, and the UWB positioning accuracy is improved by about 45.59%. The improved positioning error is less than 10 cm. The research results indicate that the algorithm proposed in this article can significantly reduce ranging errors and improve positioning accuracy, which is of great significance for promoting the widespread application of high-precision indoor positioning technology in practical scenarios.