基于线性判别分析和梯度提升决策树的WLAN室内定位算法

WLAN indoor positioning algorithm based on linear discriminant analysis and gradient boosting decision tree

为了减小室内无线局域网(WLAN)环境中接收信号强度值复杂的时变特性对定位精度的影响,提出了一种基于线性判别分析和梯度提升决策树的室内定位算法。该算法利用线性判别分析(LDA)提取原始位置指纹的主要定位特征,去除冗余和噪声;接着,使用前向分布算法,将损失函数在当前模型的负梯度值作为误差的近似值,拟合一个分类回归树,并使用加法模型将生成的分类回归树线性组合,生成梯度提升决策树(GBDT)定位模型。实验结果表明,与其他室内定位算法相比,该算法的定位精度提升20%,并且减少了接入点使用个数。

In order to reduce the influence of the complex time-varying characteristic of received signal strength indication on positioning accuracy in indoor wireless local area network(WLAN) environment, a new indoor positioning algorithm based on linear discriminant analysis(LDA) and gradient boosting decision tree(GBDT) is proposed in this paper. The algorithm adopts LDA to extract the main positioning features of original location fingerprints and remove the redundant localization features and noise. Then, using the forward distribution algorithm, the negative gradient value of the loss function in current model is taken as the approximation of the error to fit a classification and regression tree. The additive model is used to linearly combine the resulting classification and regression trees and generate a GBDT positioning model. The experiment results show that compared with other indoor positioning algorithms, the positioning accuracy of the proposed algorithm is improved by more than 20%, and the algorithm also reduces the number of the required access points.