摘要
在火电厂采用超宽带(ultra-wideband,UWB)方法实现人员定位。但因火电厂复杂环境存在多径反射与非视距(non-line-of-sight,NLoS)遮挡问题,导致定位定位结果大幅跳变的情况。提出一种融合随机森林分类、神经网络误差补偿与Q学习(Q-learning)动态锚点选择的新型UWB人员定位系统,实现视距(line-of-sight,LoS)/非视距链路状态识别、测距误差修正及基站子集动态优化,无需额外增加硬件部署即可显著提升定位稳定性。仿真结果表明:采用新型UWB人员定位系统后,人员定位最大跳变误差减小至0.4 m,均方根误差(root mean squared error,RMSE)可降低85%,有效提升了火电厂复杂工业环境下人员定位的稳定性与精度。
Ultra wideband(UWB)method is used to achieve personnel positioning in thermal power plants.However,due to the complex environment of thermal power plants,issues of multipath reflection and non Line of sight(NLoS)occlusion are existed,resulting in significant fluctuations in positioning results.A novel UWB personnel positioning system,which integrates random forest classification,neural network error compensation,and Q-learning dynamic anchor point selection,is proposed so to achieve the Line of Sight(LoS)/Non Line of Sight(NLoS)link state recognition,ranging error correction and dynamic optimization of base station subsets,which significantly improve positioning stability without the need for additional hardware deployment.The simulation results show that after adopting the new UWB personnel positioning system,the maximum jump error of personnel positioning is reduced to 0.4 m,and the Root mean squared error(RMSE)can be reduced by 85%,which effectively improve the stability and accuracy of personnel positioning in complex industrial environments of thermal power plants.
作者
廖嘉亨
吕智嘉
王佳杨
尹民
杨沛豪
LIAO Jiaheng;LYU Zhijia;WANG Jiayang;YIN Ming;YANG Peihao(Guoneng(Lianjiang)Port Power Co.,Ltd.,Fuzhou 350512,China;Xi’an Thermal Power Research Institute Co.,Ltd.,Xi’an,710054,China)
出处
《电力电容器与无功补偿》
2025年第6期97-103,共7页
Power Capacitor & Reactive Power Compensation
基金
陕西省自然科学基础研究计划(2024JC-YBMS-419)。
