基于LS-SVM步长智能判识的行人室内定位方法研究  

Research on Pedestrian Indoor Positioning Method Based on LS-SVM Step Size Intelligent Identification

作　　者：杨海 刘杰[1] YANG Hai;LIU Jie(School of Mechanical and Electrical Engineering,Southwest Petroleum University,Chengdu Sichuan 610500,China)

机构地区：[1]西南石油大学机电工程学院,四川成都610500

出　　处：《传感技术学报》2025年第1期135-142,共8页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金项目(52204013)。

摘　　要：针对行人航迹推算中步长误差累计问题,提出了一种基于最小二乘支持向量机步长智能判识行人室内定位方法,首先,将三轴加速度传感器固定在实验对象脚背上,采集行走时的加速度信号,再对加速度信号进行特征提取、特征选择,将选取的加速度信号特征集分为训练集和测试集,训练集用于训练最小二乘支持向量机步长智能判识模型,测试集用于测试模型精度,识别行人步长,结合方位角和步点解算出行人的轨迹。最后实验结果表明:步长识别准确率为96.72%,行人正常走720 m平均定位误差为0.36 m,最大定位误差为0.82 m,室内定位精度能达到95.70%。所提方法避免了步长的繁琐解算和误差累计,定位快速且精度可靠。Targeting at the accumulation of step error in pedestrian track calculation,an intelligent identification of pedestrian indoor posi-tioning method based on the least squares support vector machine is proposed.First,the three-axis acceleration sensor is fixed on the back of the subjects foot to collect the acceleration signal when walking.The acceleration signal is subjected to feature extraction and feature selection,and the selected acceleration signal feature set is divided into training set and test set,the training set is used to train the least squares support vector machine step length intelligent discrimination model,and the test set is used to test the accuracy of the model,iden-tify the pedestrian step length,and solve the trajectory of the traveler by combining azimuth and step point.The final experimental results show that the accuracy of step recognition is 96.72%,the average positioning error of pedestrians walking 720 m is 0.36 m,the maximum positioning error is 0.82 m,and the indoor positioning accuracy can reach 95.70%.The proposed method avoids the tedious calculation of step length and error accumulation,the positioning is fast and the accuracy is reliable.

关 键 词：室内定位 三轴加速度 最小二乘支持向量机 特征提取 

分 类 号：TN92[电子电信—通信与信息系统] TP181[电子电信—信息与通信工程] TP212.9[自动化与计算机技术—控制理论与控制工程]