融合神经网络和指纹的可见光定位算法研究

Research on Visible Light Location Algorithm Based on Neural Network and Fingerprint

摘要:

针对人们对室内定位需求的不断提高,以及现有室内定位算法定位精度不高等问题,提出了一种融合神经网络和可见光指纹的室内高精度定位算法。该算法利用反向传播神经网络(BPNN)确定待测目标的粗略位置,并以其预测坐标和最大误差作为约束条件,进行指纹匹配以确定待测目标精确位置。仿真结果表明,该算法平均定位误差为1.5cm,具有一定的应用价值。

Abstract:

Aiming at the continuous requirements on indoor positioning and the low positioning accuracy of existing indoor positioning methods, an indoor high-precision positioning algorithm combining neural network and visible light fingerprint was proposed. The algorithm uses the back propagation neural network (BPNN) to determine the rough position of the target, and uses its predicted coordinates and maximum error as constraints to perform fingerprint matching to determine the precise position of the target. The simulation results demonstrate that the average positioning error of the proposed algorithm can reach 1.5cm.