论文部分内容阅读
随着低功耗蓝牙的更新换代(BLFA.2),蓝牙技术在智能家居、智能设备开发中日显突出。本文针对蓝牙信号多径、易衰弱、穿透能力低等缺陷,采用支持向量机(SVM)神经网络算法来实现系统的室内定位,通过RBF径相基函数为核函数(高斯核)可控的将低维线性不可分数据映射到高维模型中,计算超平面并完成数据分类;基于空间的不可瞬变性,采用反馈滤波对验证结果进行筛选,实现精确定位。实验结果中表明算法定位精度达到1m,与传统算法相比有效的提高了效率与精度。
With the replacement of Bluetooth Low Energy (BLFA.2), Bluetooth technology has become increasingly prominent in smart home and smart device development. In order to overcome the shortcomings of multi-path, weakness and low penetrability of Bluetooth signals, this paper uses support vector machine (SVM) neural network algorithm to realize the indoor positioning of the system. The controllable RBF basis function is the kernel function (Gaussian kernel) The low-dimensional linear inseparable data is mapped into the high-dimensional model, the hyperplane is calculated and the data classification is completed. Based on the immeasurable space, the feedback filtering is used to filter the verification results to achieve precise positioning. Experimental results show that the algorithm can achieve the positioning accuracy of 1m, which improves the efficiency and accuracy compared with the traditional algorithm.