基于微惯性单元和可见光通信的室内组合定位关键技术研究

Compared to existing wireless indoor positioning technologies, Visible Light Communication (VLC) positioning technology has several features, such as immune to electromagnetic interference, high positioning accuracy, and low costs for deployment and maintenance. Therefore, VLC positioning has great advantage to achieve centimeter-level accuracy in industrial manufacturing sites with complex indoor electromagnetic field. However, there are three difficulties for VLC positioning to achieve this goal in the real-world: detecting and modelling light noise and interference, positioning in the environment with lighted signals blocked, and accurate attitude estimation for VLC receiver. In this research, Allan Variance and interference monitoring are utilized to analyze the temporal and spatial distribution of the noise and interference in VLC positioning systems and further to eliminate their effects. To overcome the latter two difficulties, an integrated VLC/MEMS system based on attitude/positioning dual-filter dual-feedback architecture is proposed to accurately estimate and eliminate the positioning errors caused by attitude change of the VLC receiver and blocked visible light signals. It is expected to establish a pervasive, accurate, and robust indoor navigation by fusing VLC localization, pedestrian dead reckoning, and multisensor attitude, which is suitable for complex indoor environments and achieves the accuracy of 30 cm.

可见光通信（Visible Light Communication, VLC）定位技术相比于现有的无线室内定位技术具有不受电磁信号干扰、定位精度高、建设成本和维护费用低的特点，在工业制造现场复杂的室内电磁环境下实现厘米级精确定位具有很大优势。但VLC定位还有三个方面的问题有待解决：光噪声干扰情况复杂难以建模、室内遮挡情况复杂难以预测和接收设备的使用姿态复杂多变难以控制和准确估计。本课题针对第一个问题拟采用基于Allan方差和干扰监测的光噪声估计方法分析VLC光噪声在时间和空间上的分布特性进行实时补偿来抑制光噪声干扰；针对后两个问题拟设计基于VLC和微惯性传感器组合的测姿定位双滤波器双反馈架构，准确估计和消除接收器姿态变化和可见光被遮挡造成的误差。预期建立一套融合VLC位置、行人航位推算位置和多传感器姿态的室内精确定位系统，能够适应复杂的实际室内环境，达到30厘米的定位精度。

可见光通信（Visible Light Communication, VLC）定位技术相比于现有的无线室内定位技术具有不受电磁信号干扰、定位精度高、建设成本和维护费用低的特点，在工业制造现场复杂的室内电磁环境下实现厘米级精确定位具有很大优势。但可见光定位（Visible Light Positioning, VLP）存在光噪声干扰情况复杂难以建模的 问题。本课题研究了不同类型噪声和外界光源干扰对可见光定位（Visible Light Positioning, VLP）系统的影响，探索了基于Allan方差的噪声估计和降噪定位方法、基于光敏二极管（Photodiode，PD）和图像传感器的组合定位方法、基于集合卡尔曼滤波的高精度融合定位方法等关键技术，建立基于CanLES的可见光定位系统的仿真验证平台，实现了适用于复杂实际场景的可见光通信与惯性传感器和图像传感器融合的室内定位系统。仿真结果和实际环境测试均表明，课题所研究的可见光定位系统降噪和抗干扰等关键技术可以有效提高可见光定位系统在复杂场景下的定位精度，平均定位精度可以达到0.09m，相比其他定位方案提高1倍以上，有着显著优势。