【摘要】 针对在煤矿井下传统Wi-Fi测距定位技术受非视距传播影响定位精度降低问题，提出一种基于改进粒子群算法优化最小二乘支持向量机的5G井下高精度定位方法。为实现视距信号和非视距信号分类，采用自适应惯性因子和变异过程的改进粒子群算法对最小二乘支持向量机参数寻优以提高训练模型的分类准确度；使用训练得到的分类模型对信号分类并使用视距信号进行基于到达时间的测距定位，提高定位精度。仿真结果表明，在实际场景中所提算法的分类准确度较支持向量机提高5.15%；且较传统粒子群优化最小二乘支持向量机定位误差小于3 m的概率提高5.66%，误差小于1 m的概率提高7.83%。更多还原

【Abstract】 In order to address the problem of reduced positioning accuracy of traditional Wi-Fi positioning technology due to non-line-of-sight propagation in underground coal mines. A combination method which based on improved Particle Swarm Optimization optimizes Least Squares Support Vector Machines is proposed for high positioning accuracy in 5G underground mines. To classify both line-of-sight and nonline-of-sight signals,the Least Squares Support Vector Machines is optimized using an improved PSO with an adaptive inertia factor and a mutation process to improve the classification accuracy of the training model;The classification model is used to classify the signals and the line-of-sight signals are used to improve the localization accuracy based on the arrival time. The simulation results show that the classification accuracy of the proposed method in the actual coal mine scenario is improved by 5.15%,and the probability of localization error less than three meters and one meter is improved by 5.66% and 7.83% compared with the traditional Particle Swarm Optimization LS-SVM.更多还原