基于PSO-BP算法的无线传感器网络定位优化

发布时间：2018-03-09 15:46

本文选题：无线传感器网络　切入点：定位算法　出处：《电讯技术》2017年02期 　论文类型：期刊论文

【摘要】：在研究现有定位算法的基础上,针对基于接收信号强度指示(RSSI)定位模型中的参数易受环境影响等问题,提出了一种新型的粒子群优化(PSO)算法与后向传播(BP)神经网络相结合的算法。BP网络算法权值的修正依赖于非线性梯度值,易形成局部极值,同时学习次数较多,需先通过粒子群算法进行优化。为了提高定位精度,首先采用速度常量法滤波处理,然后通过改进的混合优化算法对BP神经网络初始权值和阈值进行优化,并分析算法的性能。试验中隐层节点个数采用试错法,从12到19变化,以确定合适数目。实验结果表明,与一般加权算法和传统BP算法相比,改进的混合优化算法可大幅改善测距误差对定位误差的影响,同时可使25 m内最小定位误差小于0.27 m。
[Abstract]:On the basis of studying the existing localization algorithms, aiming at the problem that the parameters in the RSSI location model based on the received signal strength indicator are vulnerable to the environmental influence and so on, A new Particle Swarm Optimization (PSO) algorithm combined with backward Propagation (BPN) neural network is proposed. The weight correction of BP network algorithm depends on the nonlinear gradient value, which is easy to form the local extremum and has more learning times. In order to improve the positioning accuracy, the speed constant method is first used to filter, and then the initial weight and threshold of BP neural network are optimized by the improved hybrid optimization algorithm. The performance of the algorithm is analyzed. In the experiment, the number of hidden layer nodes varies from 12 to 19 to determine the appropriate number. The experimental results show that, compared with the general weighted algorithm and the traditional BP algorithm, the number of hidden layer nodes in the experiment is better than that of the traditional BP algorithm. The improved hybrid optimization algorithm can greatly improve the influence of the ranging error on the positioning error and make the minimum positioning error less than 0.27 m within 25 m.
【作者单位】：贵州大学现代制造技术教育部重点实验室;中国海洋大学工程学院;国网山东省电力公司潍坊供电公司;
【基金】：贵州省科技支撑计划(黔科合GZ字[2015]3034) 国家自然科学基金资助项目(51475097) 国家科技支撑计划(2014BAH05F01) 贵州省科技基金项目(黔科合J字[2015]2043) 贵州省基础研究重大专项(黔科合JZ字[2014]2001)
【分类号】：TN929.5;TP212.9

【相似文献】