Research On Location Technology Of Search And Rescue Personnel In Sheltered In Sheltered Space Based On UWB Signal

In recent years,a series of safety accidents such as fires and earthquakes have occurred frequently in China.Search and rescue personnel often go to various sheltered spaces to carry out rescue.However,in the process of rescue,search and rescue personnel often greatly reduce the efficiency of rescue because they cannot determine their exact location.Therefore,how to carry out high-precision positioning indoors becomes a key issue.Ultra Wide Band(UWB)has great potential in building highprecision indoor positioning system due to its large bandwidth and strong penetration characteristics,which has been widely studied by domestic and foreign scholars.This thesis further explores the field of UWB non-line of sight(NLOS)recognition and location algorithms.The main contents and innovations are as follows:1.Aiming at the problem of low NLOS signal recognition rate in occlusion scenarios,a three-channel convolutional neural network combined with UWB signal recognition model of bidirectional long shortterm memory network(TC CNN-BiLSTM)is designed.Because the channel impulse response(CIR)of the UWB signal has rich characteristic information and there is temporal correlation between the front and back sampling points.Based on this,this thesis takes the CIR sequence and its Fourier transform real and imaginary part sequence as the three-channel input.We use CNN to extract the features and BiLSTM to carry out bidirectional temporal analysis on the sequence,and finally complete the NLOS/LOS signal classification of the UWB.The simulation results show that the NLOS/LOS classification accuracy of TC CNN-BiLSTM can reach 85.71%.Compared with the LSTM and CNN-LSTM models with higher classification accuracy in the existing research,the classification accuracy has increased by 10.2%and 4.7%respectively.2.Due to the complex and changeable search and rescue scenarios,the traditional UWB indoor positioning algorithm often causes problems such as poor positioning accuracy and underpositioning,and in order to solve this problem,this thesis designs a UWB directional positioning technology based on genetic algorithm to optimize rotational TOF.This method uses the search and rescue personnel to carry the inertial navigation system and hold the UWB equipment to rotate for one circle to obtain the direction angle change sequence and the TOF ranging sequence between the single base station.In order to fit the TOF ranging sequence into a standard sine curve,It is abstracted as a least-squares fitting problem.At the same time,in order to avoid falling into the local optimal solution,the genetic algorithm is used to optimize it.Finally,the rotation angle corresponding to the minimum value of the standard sine fitting curve is unified with the TOF ranging value at this time,so as to complete the search and rescue personnel position calculation with low cost and simple operation.After simulation experiments,the average directional accuracy deviation of the proposed method is 11.67°,which is 6.48° higher than the legal accuracy of template matching.In the two typical masking scenarios of underground tunnel and underground garage,the actual positioning accuracy of the algorithm is 27.44cm and 20.11cm,respectively,which has obvious performance improvement compared with the trilateral positioning algorithm in the NLOS/LOS mixed environment.