ISSN: 1673-825X    Imprint: Chongqing University of Posts and Telecommunications Journal
Volume listings
  • Current volume
  • Search by keywords:
  • Search by 
  • Keywords
  • To
  • Journal archive
Related Links
Fingerprint location using sparse fingerprint acquisition and improved WKNN
Received:May 21, 2018  Revised:April 30, 2019
中文关键词:室内指纹定位  高斯过程回归  超参数  共栖生物搜索  卡方距离
英文关键词:indoor fingerprint location  Gaussian process regression  hyper parameter  symbiotic organisms search  chi-square distance
Author NameAffiliationE-mail
LI Xinchun School of Electronic and Information Engineering,Liaoning Technical University,Huludao 125105, P. R. China 
WANG Huan College of Graduate Studies,Liaoning Technical University,Huludao 125105, P. R. China 
Hits: 189
Download times: 73
      为解决位置指纹定位算法中指纹采集工作量大、定位精度低的问题,提出一种基于稀疏指纹采集和改进加权K最近邻(weighted k-nearest neighbor,WKNN)的定位算法。稀疏选定参考点并采集来自各接入点(access point,AP)的接收信号强度(received signal strength,RSS),根据容错四分位法对采集的RSS进行异常值预处理;利用经过预处理的指纹数据训练高斯过程回归(Gaussian process regression,GPR)模型,通过共栖生物搜索算法(symbiotic organisms search,SOS)求取模型最优超参数以提高模型的泛化能力,进而预测定位区域内非参考点的RSS;由有限参考点数据通过SOS-GPR模型的训练与预测生成密集位置指纹库,结合由卡方距离和AP加权改进的WKNN算法完成仿真验证。实验结果表明,在保证定位精度的前提下,稀疏指纹采集法较传统全采集法减少50%的采集工作量;与原WKNN算法和M-KWNN算法相比,提出的WKNN算法有效提高了定位精度。
      To solve the problem on heavy fingerprint acquisition workload and low positioning accuracy of fingerprint location, this paper proposes a method of using sparse fingerprint acquisition and improved WKNN algorithm to locate. Firstly, it selects reference nodes sparsely, collects received signal strength (RSS) from access points (APs) and processes outliers by fault-tolerant quartiles method. Secondly, it uses the preprocessed fingerprint data to train Gaussian process regression (GPR) model, optimizes the model’s hyper parameters by symbiotic organisms search (SOS) algorithm, and generates the fingerprint prediction model based on SOS-GPR to predict non-reference nodes’ RSS. Finally, a dense location fingerprint database is constructed by limited reference nodes, and combines with the improved WKNN algorithm by chi-square distance and AP weighted to complete positioning simulation. Compared with traditional database construction method WKNN and M-WKNN algorithm, experimental results show that the proposed algorithm can not only reduce acquisition workload by 50% under the premise of ensuring positioning accuracy, but also improve positioning accuracy effectively.
HTML    View PDF   View/Add Comment  Download reader