Fingerprint Indoor Positioning Method Based on Siamese Network

Abstract

The invention discloses a fingerprint indoor positioning method based on a twin network, which effectively improve the positioning accuracy of an indoor positioning environment where samples are difficult to collect. According to the technical scheme, the method comprises the steps of 1) constructing a fingerprint indoor positioning system; 2) dividing an indoor positioning area; 3) estimating channel state information of each reference point; 4) calculating the fingerprint of each reference point; 5) constructing a training set and a verification set; 6) training the twin network model; 7) constructing a channel state information sample pair set of a to-be-positioned point; and 8) obtaining the position of a to-be-positioned point. Based on the fingerprint indoor positioning method, the problem that when a training set is small, a traditional classification method is difficult to train an accurate classification model is considered, sample pairs are constructed, the size of the training set of the twin network is effectively increased, and the trained model has higher classification capacity. The method is used in an indoor environment where samples are difficult to collect, and is suitable for indoor rescue, personnel positioning in large-scale places and other scenes.

Description

technical field [0001] The invention belongs to the field of communication technology, in particular to a fingerprint indoor positioning method, in particular to a twin network-based fingerprint indoor positioning method. It is applicable to indoor scenes with fewer obstacles, especially in indoor scenes where it is difficult to collect information, and can provide users with accurate positioning and a good user experience. Background technique [0002] With the development of communication technology, location-based service has become a new mobile Internet industry and has a good development prospect. Therefore, the need to quickly and accurately obtain location information of a mobile terminal has become increasingly urgent. Positioning information can also be used to support location-based services and improve network management, improving the quality of location services and network performance. Therefore, the positioning technology and its related positioning system t...

AI Technical Summary

Problems solved by technology

This invention uses wavelet transform to process the data of reference points, and then uses convolutional neural network to classify, which can effectively improve the positioning accuracy. Therefore, this method is only suitable for indoor positioning scenarios where each reference point has a large number of sample sets, and the positioning accuracy is lower in indoor positioning environments where it is difficult to collect samples. Low

Images