论文部分内容阅读
针对室内环境基于RSSI定位不稳定问题,提出了以几何信息改进基于指纹库的KNN定位算法。根据室内几何布局建立了聚类指纹库,提出了表征点位几何特性的点散发性强度(geometric strength of sporadic,GSS)概念。利用最邻近样本点的GSS判别移动终端所在参考点RP控制网结构以动态选择KNN关键参数K,构建最佳多边形为约束准则自适应选取后K-1个邻近点,建立了基于几何聚类指纹库的约束加权KNN室内定位模型。结果表明,改进后定位模型可以更好地估计终端位置信息,其中几何聚类指纹库是改善定位准确性的关键,约束KNN能够有效地提高室内定位精度。
Aiming at the instability of indoor environment based on RSSI positioning, a new algorithm of KNN localization based on fingerprint database is proposed. According to the geometrical layout of the interior, a fingerprint database was established and the concept of geometric strength of sporadic (GSS) was proposed. The nearest neighbor point (GSS) is used to determine the reference point of the mobile terminal RP control network structure to dynamically select the KNN key parameter K, and the optimal polygon is constructed as the constraint criterion to adaptively select K-1 neighboring points. A geometric clustering fingerprint Constraint weighted KNN indoor location model for library. The results show that the improved localization model can better estimate the terminal location information. The geometric clustering fingerprint database is the key to improve the positioning accuracy. The restrained KNN can effectively improve the indoor positioning accuracy.