联合光源与视觉信息的高精度可见光定位理论与技术研究

Visible light positioning (VLP) is a novel indoor localization technology based on light-emitting diode (LED) beacon. Due to the limitation of localization principle and the insufficiency of the utilized information, current VLP approaches still suffer from problems including the orientation limitation of receivers, the insufficiency of coverage rate, and the unreliability localization for users in moving. Based on the joint utilization of the light information such as signal strength and visual information such as the positions of LEDs, this project proposes novel joint localization technology that combines light information and visual information, which can achieve high localization accuracy while overcoming the aforementioned challenges. To solve the receiver orientation limitation problem, the project proposes a novel joint VLP algorithm based on multiple view geometry. To solve the coverage rate insufficiency problem, the project proposes a novel adaptive multi-mode joint VLP algorithm. To solve the unreliability localization problem in mobile positioning, the project proposes a novel joint VLP algorithm based on dynamic switches of LEDs. Based on these joint VLP algorithms, the project establishes the VLP prototype to prove the feasibility of the proposed approaches and analyze the localization performance. The project investigates the VLP technology that utilizes multi-information sources, which is in line with the instructions of the government’s 13th Five-Year Plan that developing new localization technologies for multi-information sources, and will definitely support the fulfillment of national strategic tasks such as seamless indoor and outdoor localizations and smart space-time.

可见光定位是一种基于LED信标的新型室内定位技术。由于定位原理的局限和信源的单一，现有可见光定位技术还存在接收机朝向受约束、定位覆盖率不足以及移动中定位可靠性受限等挑战。课题通过联合可见光信号强度等光源信息与LED信标空间位置等视觉信息，创新性地提出联合光源与视觉信息的可见光定位技术，在有效保障高定位精度的同时克服上述难点。课题针对接收机朝向受约束问题，研究新型基于多视图几何的联合可见光定位理论与方法；针对可见光定位覆盖率不足问题，研究新型自适应多模的联合可见光定位理论与方法；针对移动中定位可靠性受限问题，研究新型LED动态切换的联合可见光定位理论与方法；在此基础上，搭建可见光定位平台，证明所提方法的可行性并验证定位性能。课题探索利用多源信息的新型可见光定位技术，符合我国“十三五”规划开展面向多信号源的新型定位技术的指导意见，将有力支撑未来室内外无缝定位与智慧时空等国家战略任务的顺利实现。

项目挖掘了可见光定位和视觉定位的互通性，围绕如何解决可见光定位在实际应用中所面临的挑战展开。项目具体针对接收机朝向受约束问题，联合利用视觉、强度信息，提出了新型基于多视图几何的联合可见光定位算法；针对可见光定位的室内覆盖率不足问题，建模基于直线、曲线特征的投影模型，提出了新型自适应多模的联合可见光定位算法；针对接收机移动过程中的可靠定位问题，研究新型LED动态切换的联合可见光定位算法；最后搭建了联合光源与视觉信息的可见光定位平台进行性能分析及验证，实测表明项目所提定位算法在接收端朝向定位精度10厘米范围内，定位响应时延在0.13s内，验证了项目所提可见光定位算法在实际应用中的性能优势。.项目的研究充分挖掘可见光的光源强度信息和视觉空间信息用于定位，并且巧妙地联合PD和图像传感器实现多源信息接收与处理，有效促进了可见光定位与视觉定位的联合和增强。这一解决思路增加了空间这一定位自由度，在满足低成本前提下进一步提升定位精度成为可能，也为包含可见光定位的室内定位提供了指导意义。.研究成果体现在发表高水平学术论文21篇，其中在IEEE Transactions on Wireless Communications、IEEE Transactions on Communications等通信领域顶级期刊发表SCI检索论文10篇，EI检索论文11篇，申请发明专利7项并获得2项授权，培养博士和硕士研究生14人等。