基于BP神经网络和APIT室内定位算法的研究与实现

Implementation of Indoor Localization Algorithm based on BP Neural Network and APIT

随着国内外无线通信技术的发展,无线定位技术成为必不可少的一项技术。现有的室外定位技术已日渐成熟,而室内定位技术由于定位精度不够等原因,还有很大发展空间。在分析无线通信模型的基础上,针对室内无线信号传播模型中参数A和路径损耗指数n依经验判断而定导致的定位误差太大问题,利用BP神经网络拟合RSSI-d非线性函数关系,大大缩小了定位误差。在得到对应的距离值后,利用改进的APIT室内定位算法确定盲节点的坐标,并在Zigbee开发平台对该算法进行实现,验证了算法的可行性,减小了定位误差。

With the development of wireless communication technology both at home and abroad, the wireless location becomes an essential technology. The existing outdoor positioning technology, although is getting mature, still has plenty of room for growth due to its not-enough positioning technology because of the lack of positioning accuracy. Based on the analysis of wireless communication model and for the problem that the empirical determination of parameter A and path loss index N in the indoor wireless signal propagation model would result in too large location error, BP neutral network is used to fit the nonlinear function of RSSI-d,thus greatly reducing the positioning error. After the acquistion of corresponding distance value and with the improved APIT indoor location algorithm, the coordinates of blind nodes could be determined. The implementation on the Zigbee development platform indicates the feasibility and effectiveness of the algorithm.