基于蚁群算法和BP神经网络的信道分配策略的研究  

Research on Channel Allocation Strategy Based on Ant Colony Algorithm and BP Neural Network

作　　者：翟学明[1] 王佳[1] 李金泽[1] 

机构地区：[1]华北电力大学(保定)控制与计算机工程学院,河北保定071003

出　　处：《传感技术学报》2016年第3期445-450,共6页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金项目(60974125)

摘　　要：研究无线传感器网络信道分配策略的主要目标是提高网络吞吐量和容量,减小网络的传输时延,最大限度的利用有限的网络带宽资源。多信道MAC协议的应用,可以有效地提高网络通信的可靠性和吞吐量,以及解决由于信道受干扰而造成的网络瘫痪等问题。根据无线传感器网络多信道的特点提出了一种基于蚁群算法的动态反馈负载均衡信道分配策略。本策略首先应用BP神经网络对信道负载情况进行预测,然后通过基于蚁群算法的负载均衡算法对信道进行筛选,最后利用最大离散化算法进行信道分配。在NS2平台下对所设计的协议进行了仿真实现,并与应用最为广泛的多信道MMAC协议以及SMAC进行了对比分析。根据仿真结果可知,本文设计的MAC协议在网络吞吐量、网络传输时延等性能方面比MMAC协议及SMAC都有了很大程度的提升。可以有效减小网络传输时延,提高网络吞吐量和抗干扰能力。The main purpose of studying channel allocation strategy of wireless sensor network is to improve network throughput and capacity, reduce transmission delay of the network, and use the limited network bandwidth resources efficiently. Application of multi-channel MAC protocol can solve the problem of network paralysis caused by channel interference. Due to the characteristics of the wireless sensor network, this article proposed a dynamicfeedback load balancing channel allocation strategy based onant colony algorithm. This strategy applied BP neural network to predict the channels＇ load, then filtered the channels by using the load-balancing algorithm based on ant colony, finally allocated the channel sequences by using the maximum discrete channel allocation algorithm. Finally, this protocol wassimulated and implemented on NS2 platform, and it was also compared with the MMAC protocol and the SMAC protocol. According to the simulation, the MAC protocol we proposed in this paper performed better than the MMAC protocol and the SMAC protocol in terms of network throughput, network delay to a large degree of improvement. It can effectively reduce the network delay, increase network throughput and enhance the anti-jamming capability.

关 键 词：无线传感器网络 信道分配机制 蚁群算法 BP神经网络 最大离散化 NS2仿真 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]