基于GM(1,1)模型的室内定位指纹库误差修正方法

Error Correction Method of Indoor Location Fingerprint Based on GM(1,1) Model

环境噪声干扰造成的室内蓝牙指纹库误差,直接对室内定位精度产生不利影响。通过构建室内蓝牙指纹库的GM(1,1)模型,对位置指纹库的误差进行修正,为在线阶段的匹配定位算法提供了更优质的指纹库数据。分别将GM(1,1)模型的残差修正法和Kalman滤波去噪法优化后的两种指纹库运用于K近邻、加权近邻、贝叶斯匹配算法中,仿真结果表明,GM(1,1)模型的残差修正法所获得的定位精度更高。

The error of indoor Bluetooth fingerprint database caused by environmental noise interference has a direct impact on the accuracy of indoor positioning.In this paper,the GM（1,1）model of indoor fingerprint database is used to modify the error of position fingerprint database,the matching positioning algorithm provides better quality fingerprint database data.The GM（1,1）models residual error correction method is applied to the K nearest neighbors algorithm,the weighted K nearest neighbors algorithm and Bayes localization algorithm respectively,the Kalman filter denoising method as above.The simulation results show that the GM（1,1）model＇s residual error correction method obtains higher positioning accuracy.