In this paper, a novel three-dimensional coordinate positioning algorithm based on factor graph is proposed to improve the measurement accuracy of the indoor global positioning system (iGPS) under large scale conditions. Different from the traditional iGPS positioning algorithm based on the least squares estimation, which has the problems of fixed solution model, low confidence results and poor stability, this proposed algorithm utilized Bayesian filtering to solve the coordinate positioning problem. Aiming at the character of plug and play for iGPS, a factor graph model based on Bayesian network is built, and then a sum product algorithm is used to convert the fixed model into the form of the product of every node, which reduces the independence of measurement information, and improves the confidence of the results. Furthermore, to further improve the positioning accuracy of the algorithm, an idea of maximum posterior estimation is integrated into the proposed algorithm, which enhances the stability of the algorithm at the same time. In order to verify the effectiveness of the proposed algorithm, a series of simulation and prototype tests have been carried out. The results show that compared with the traditional positioning algorithm based on least squares estimation, the accuracy of the proposed algorithm is improved by about 50%, and the positioning accuracy can achieve 0.3mm within a range of 10 meters, which realizes a high-precision measurement under large scale conditions.

中文翻译：

---

一种基于因子图的三维坐标定位新算法


本文提出一种基于因子图的三维坐标定位算法，以提高室内全球定位系统（iGPS）在大尺度条件下的测量精度。与传统的基于最小二乘估计的iGPS定位算法存在求解模型固定、结果置信度低、稳定性差的问题不同，该算法利用贝叶斯滤波来解决坐标定位问题。针对iGPS即插即用的特点，建立了基于贝叶斯网络的因子图模型，然后利用和积算法将固定模型转化为各节点乘积的形式，降低了iGPS的独立性。测量信息，提高结果的置信度。此外，为了进一步提高算法的定位精度，算法中融入了最大后验估计的思想，同时增强了算法的稳定性。为了验证所提算法的有效性，进行了一系列仿真和原型测试。结果表明，与传统的基于最小二乘估计的定位算法相比，所提算法的精度提高了约50%，在10米范围内定位精度可达到0.3mm，实现了高精度定位。大尺度条件下的测量。