由于卫星信号无法穿墙透壁，因此全球卫星导航系统无法提供精确的室内定位服务。高精度室内定位技术是物联网、人工智能应用的基础核心技术之一。然而大多数已有的厘米级光定位算法要求接收机垂直于水平面，这对于手持设备和非平坦路面并不现实。本项目将研究一种和接收机方向角无关的，基于光信号到达角度差（ADOA）的室内定位新方法，拟实现：1）建立ADOA光定位的理论。基于矩阵扰动分析理论推导ADOA定位系统的闭式误差上下界和渐近误差公式；2）设计高速、高精度的ADOA光定位算法。以定位误差公式为准则，基于并行计算、自适应可行域和分层搜索技术，在计算机平台上实现平均每秒10次以上的定位速度和5厘米以内的定位精度；3）优化面向ADOA估计的光传感器阵列。基于特征值分析和最优化方法，求解最佳传感器倾角组合，从而最小化角度估计误差。该研究将为高精度、高速度的ADOA室内光定位方法提供理论依据和实践指导。

Due to the high attenuation of satellite signals in indoor scenarios, Global Navigation Satellite Systems cannot provide accurate indoor positioning services. Accurate indoor positioning technology is one of the key technologies for the internet of things and artificial intelligence applications. However, Most of existing centimeter-level optical positioning algorithms require the receiver to be vertically placed, which is inconvenient for handheld devices or rough ground. In this project, we propose a new positioning scheme based on angle difference of arrival (ADOA), which is irrelative to the orientation of the receiver. We are going to realize the following targets. 1) Lay foundation for the theory of ADOA positioning, deriving closed-form approximation and bounds for the error of ADOA positioning schemes based on matrix perturbation analysis. 2) Propose fast and accurate ADOA optical positioning algorithms based on the error expressions and techniques including parallel computing, adaptive feasible regions and hierarchical searching, and realize a positioning speed of ten times per second and average positioning accuracy less than 5 centimeters on a computer. 3) Propose an optimization framework for the optical sensor arrays for ADOA measurement. Based on the analysis of eigenvalues and optimization techniques, we are going to solve the best combination of sensor angles in order to minimize the angle estimation error. The research will provide theoretical foundation and practical suggestions for the accurate and high-speed ADOA positioning systems.