视距与非视距共存场景室内无线定位方法研究

凭借极低部署成本，WiFi室内定位已成为无线定位领域新兴热点研究方向。复杂室内环境造成节点之间无线信号视距与非视距传播共存。针对现有方法在复杂室内环境中定位精确度较低的问题，本项目拟探索适用于视距/非视距共存场景的基于信道状态信息的定位方法。其核心思想是通过为视距和非视距场景分别建立无线信道模型和距离/角度无偏估计模型，推算出视距与非视距混合场景定位误差理论下界；通过在信道状态信息幅度与相位组合特征的基础上建立机器学习模型，提出视距和非视距智能区分方法，构建距离和角度测量误差消除方法；通过自回归模型、非线性滤波、自适应贝叶斯和神经网络等信息融合技术，提出视距与非视距测量融合定位方法和指纹与测量模型融合定位方法。本项目是对室内无线定位理论与技术研究的进一步深化，为解决高精度低成本室内无线定位问题探索一种新的研究手段，为基于信道状态信息的室内定位方法未来能真正走向实用化奠定良好理论和技术基础。

Due to low deployment cost, WiFi based indoor localization has become an emerging hot research topic on indoor wireless localization. Complex indoor environment causes the coexistence of line-of-sight (LOS) and non line-of-sight (NLOS) propagation of wireless signals. Considering the accuracy problem of CSI-based localization in hybrid line-of-sight (LOS) and NLOS scenarios, this proposal focuses on channel state information (CSI) based indoor localization schemes that can be applied for LOS/NLOS coexistence scenarios. The main principle is explained as follows. Firstly the lower bound of localization error under LOS/NLOS coexistence scenario will be derived based on the modeling of wireless channel and unbiased distance/angle measurement. Secondly, LOS/NLOS identification method and distance/angle measurement error mitigation method will be proposed through machine learning techniques which take the combination of the amplitude and phase of CSI as the input. Thirdly, LOS/NLOS measurement fusion localization method and fingerprinting/measurement model fusion localization method will be proposed based on information fusion techniques such as autoregressive model, non-linear filtering, adaptive Bayes decision and neural network. This proposal has deep research on indoor wireless localization theory and technique, provides a new research approach for high-accuracy low-cost indoor wireless localization and makes solid theoretical and technical work for the application of CSI-based indoor localization method.