基于深度学习和质点弹性优化的传感器节点定位

Sensor Nodes Locating Based on Depth Learning and Particle Elastic Optimization

为了实现无线传感器网络中传感器节点的定位,提出一种基于神经网络的深度学习定位方案。方案分别采用2个体系结构相同的神经网络集成(Neural Network Integrations,NNIs)即X-NNI和Y-NNI来估计x和y坐标,每个NNI由K个分量神经网络构成,每个分量神经网络是一个3层前馈神经网络。每个未知节点在接收到经过训练的NNIs的信息之后,使用在信息阶段收集的跳数信息作为NNIs的输入来估计其位置;还提出一种利用相邻信标节点和未知节点的位置信息的增强型质点弹性优化位置估计算法来进一步提高系统的整体定位性能。仿真实验结果表明,提出的定位方案相比于其他无距离定位算法以及基于单神经网络的定位算法在平均定位误差方面有显著改善,而且具有良好的实用性和稳定性。

In order to realize the location of sensor nodes in wireless sensor networks,a deep learning location scheme based on neural network is proposed.In this scheme,two Neural Network Integrations(NNIs)with the same architectures X-NNI and Y-NNI are used to estimate the x and y coordinates respectively.Each NNI consists of K component neural network where each component neural network is a three-layer feed forward neural network.After receiving the information of the trained NNIs,each unknown node’s location is estimated with the hop-count information gathered in the info phase as input to the NNIs.An enhanced particle elastic optimization’s location estimation algorithm that utilizes the location information of both the neighboring beacon and unknown nodes is proposed to further improve the overall location performance of the system.The simulated results show that the proposed location scheme has a significant improvement on the average location error compared with other range-free location algorithms as well as the location algorithm based on a single neural network,and has a good practicability and stability.