Research on UWB positioning trajectory prediction based on LSTM

In order to improve the accuracy of the neural network in predicting and positioning based on Ultra-Wideband (UWB), a method based on long and short-term memory network to predict UWB positioning data is proposed, which is comparable to ordinary neural networks that only use current information for training. Compared with the method based on long and short-term memory network prediction, the historical information is effectively used. First, the UWB bilateral two-way ranging algorithm is used to obtain the ranging value as the feature information, the Long Short-Term Memory (LSTM) model is established, and the optimal coordinates obtained by the least square method are used as the output supervision training. Finally, combined with the historical data of UWB positioning trajectory to predict the current coordinates of the test set graphics, it is verified that the method in this paper greatly improves the positioning accuracy of UWB compared with ordinary neural network prediction positioning.