针对目前井下定位系统直接测距精度不高的问题,文章提出了一种基于改进极限学习机(Extreme Learning Machine,ELM)的井下人员定位算法.文章选用基于指纹的位置匹配模型,通过引入莱维飞行策略和模拟退火机制,对秃鹰算法进行改进,采用改进秃鹰优化算法(Improved Bald Eagle Search Optimization Algorithm,IBES)优化ELM模型,旨在提高定位模型的收敛速度和全局搜索能力.仿真实验结果表明,所提IBES-ELM方法的平均定位误差为0.71 m,定位精度优于BES-ELM和K近邻(K-nearest Neighbor,KNN)及其衍生算法,验证了文章算法具有更好的预测性能和稳定性.
Downhole fingerprint localization algorithm based on improved ELM
Aiming at the problem that the direct ranging accuracy of the current downhole positioning system is not high,and a downhole personnel positioning algorithm based on improved extreme learning machine is proposed.A fingerprint-based position matching model is selected,and the bald eagle algorithm is improved by introducing the Lévy flight strategy and the simulated annealing mechanism,and the ELM model is optimized by the improved bald eagle search optimization algorithm,which is designed to improve the convergence speed of the localization model and the global search.The optimized ELM model is designed to improve the convergence speed and global search capability of the localization model.The simulation experiment results show that the average localization error of the proposed IBES-ELM method is 0.71 m,and the localization accuracy is better than that of BES-ELM and K-nearest neighbor and its derivative algorithms,which verifies that the algorithm of the article has better prediction performance and stability.
downhole positioninglocation fingerprintingimproved bald eagle algorithmextreme learning machine
NETL
NSTL
万方数据
