论文部分内容阅读
针对在Wi Fi环境下,传统的位置指纹定位算法定位精度不够高和指纹数据库构建困难的问题提出了一种基于线性加权回归(LWR)和蜂群优化的支持向量回归机(ABCSVR)的LWR-ABCSVR定位算法。该算法通过LWR在离线阶段对采集到的位置指纹数据库进行扩充;利用ABCSVR构建物理位置和RSS之间的非线性关系,并通过构建的预测模型完成定位。实验结果表明,该算法的定位精度远高于传统的几种定位算法,并且可以在一定程度上减少构建指纹数据库的工作量,是一种综合性能良好的定位算法。
Aiming at the problem that the location accuracy of traditional location fingerprinting algorithm is not high enough and the fingerprint database is difficult to build in Wi Fi environment, this paper proposes a LWR-based support vector regression machine (ABCSVR) based on linear weighted regression (LWR) ABCSVR positioning algorithm. The algorithm extends the collected position fingerprint database by LWR in offline phase. It uses ABCSVR to construct the non-linear relationship between physical location and RSS, and completes the localization through the constructed prediction model. Experimental results show that the proposed algorithm has much higher positioning accuracy than the traditional algorithms and can reduce the workload of fingerprint database construction to a certain extent.