论文部分内容阅读
针对室内定位中基于位置指纹的K近邻法采用固态K值无法得到最优定位结果的问题,提出自适应K值选择的K近邻法。算法利用相邻定位点短时间间隔内空间位置变化引起的信号强度变化规律推测运动趋势,并与不同K值的定位结果构建的空间矢量进行匹配,从而自适应地从K近邻法的不同K值中选取最优的K值。同时依据室内AP的几何布局特征划分多个矢量域内,并对定位结果进行区域改正。试验结果表明,该算法能够很好地抑制较大误差的出现,提高定位的实时性、定位精度和稳定性。
Aiming at the problem that the K-nearest neighbor method based on position fingerprint can not get the optimal positioning result in solid-state K in indoor positioning, a K-nearest neighbor method of adaptive K value selection is proposed. The algorithm uses the variation of signal intensity caused by the change of spatial position within a short time interval of adjacent positioning points to infer the movement trend and match with the space vector constructed by the positioning results of different K values to adaptively select K values from different K values Choose the best K value. At the same time, according to the geometrical layout features of indoor AP, it divides a plurality of vector domains and performs regional correction on the positioning results. The experimental results show that the proposed algorithm can restrain the occurrence of larger errors and improve the real-time positioning, positioning accuracy and stability.