室内强遮挡环境下基于近超声的位置指纹定位方法

Position fingerprint localization method under indoor strong barrier environment based on near ultrasound

针对复杂室内强遮挡、低基站部署密度的场景,研究基于近超声的位置指纹定位方法。提出了基于近超声到达时间(TOA)信息的位置指纹定位方法,并基于双曲调频信号建立高精度的TOA信息指纹库;分别结合反向传播(BP)神经网络和加权K最近邻(WKNN)提出了TOA—人工神经网络(TOA-ANN)和TOA-WKNN定位方法架构。实验结果表明:在复杂室内遮挡环境中,TOA-ANN优于TOA-WKNN的方法架构。在仅用2个声源基站的条件下,其定位误差小于42 cm的概率为90%,小于27 cm的概率为80%,定位误差小于20 cm的概率为60%。在强遮挡低基站部署密度环境中,能够满足面向智能移动终端的室内定位系统要求,具有很好的应用和推广价值。

Aimed at the scene of indoor strong barrier and low base-station deployment density,position fingerprint localization method based on near ultrasound is researched.Position fingerprint localization method based on near ultrasound time of arrival(TOA)information.TOA information fingerprint database with high precision based on hyperbolic frequency modulation signal is established.Combined with back propagation(BP)neural network and weighted K nearest neighbor(WKNN),architectures of TOA-artificial neural network(ANN)and TOA-WKNN localization methods are proposed.The experimental results show that the TOA-ANN framework is superior to TOA-WKNN framework in complex indoor scenarios.By only using 2 sound source base-station probability of the positioning error of TOA-ANN positioning framework is less than 42 cm with 90%,27cm with 80%,and 20 cm with 60%.The proposed method can meet the needs of indoor positioning system for smart mobiles in strong barrier environment with low base station deployment density,and has good application and promotion value.