| We are now living in a mobile information era.With the rapid development of wireless sensor networks and mobile intelligent terminals,indoor localization is providing support for many location-based services(LBS),and the demand for indoor location-based services is increasing.At present,fingerprint-based localization methods are commonly used in indoor localization research.Fingerprint features and localization model are two key factors affecting the localization accuracy in fingerprint localization methods.In terms of fingerprint feature,the stability of visible light intensity is high,but the discrimination of position features is low.The radio signal strength is highly distinguishable,but has strong volatility.Meanwhile,the localization models based on convolutional neural networks(CNN)can not effectively highlight important features during feature extraction.To achieve these challenges,a hybrid fingerprint indoor localization method based on ECA-CNN(ECACon-HF)is proposed in this thesis.First,this thesis use visible light intensity and received signal strength indication(RSSI)of Bluetooth low energy(BLE)to construct hybrid fingerprints,reduce the influence of instability of BLE fingerprints,and enhance the discrimination between different positions.Meanwhile,the CNN localization model is improved by the efficient channel attention(ECA).ECA can adaptively extract important information in fingerprints through cross-channel interaction strategies,suppress environmental interference in fingerprints,enhance the expression ability of hybrid fingerprint features,and make more effective use of the advantages of hybrid fingerprints.The experimental results show that the ECACon-HF method proposed in this thesis achieves the localization precision of 0.316 m,higher accuracy than on the single fingerprint.Meanwhile,based on the same fingerprint database,the proposed method outperforms other related indoor localization methods.The main contributions of this thesis are as follows:(1)This thesis constructs a hybrid feature fingerprint by combining visible light intensity and BLE RSSI.By doing so,this thesis reduce the negative impact of unstable BLE fingerprints on localization accuracy and enhance the discrimination of fingerprint features at different locations.Consequently,our method achieves a higher accuracy of fingerprint matching,resulting in a significant increase in localization precision.Notably,the acquisition of visible light intensity in this thesis does not require the use of visible light communication technology,which greatly reduces the complexity of devices.(2)This thesis propose a hybrid fingerprint indoor localization method based on ECA-CNN(ECACon-HF).The ECA module is incorporated into the CNN to construct an ECA-CNN localization model,which matches the target point with various reference points.The ECA module can adaptively extract crucial information from the fingerprint,enhance feature representation capability,and suppress irrelevant features such as environmental interference.Moreover,the ECA module explicitly models the interdependence relationship between channels through a cross-channel interaction strategy.This strategy captures the feature representation with the maximum information along the channel direction.By integrating the location correlation in visible light fingerprint and BLE fingerprint,the proposed ECACon-HF method can fully leverage the advantages of hybrid fingerprint and improve the accuracy of fingerprint feature matching.(3)The experimental environment was established,and experiments were designed and conducted to construct a hybrid feature fingerprint dataset of visible light intensity and BLE RSSI.The efficacy of the method proposed in this thesis was verified using this dataset.Finally,an indoor localization system was designed and implemented based on the ECACon-HF method proposed in this thesis,which seamlessly integrates the research content of this thesis with practical application scenarios. |
