一种新的传感器网络混合广播调度方法

由于传感器网络所使用无线信道的共享性和相互干扰, 节点间数据广播会产生资源冲突, 广播调度要解决的即是为每个节点分配到一个无冲突传输时隙, 其目标是找到最优时分复用(TDMA: time division multiple access)调度解, 使得帧长度最短而信道利用率最大. 提出基于神经网络的两阶段混合广播调度算法. 在阶段一, 使用改进的顶点着色算法来获得调度所需最短时隙数目; 在阶段二, 使用模糊Hopfield网络将节点模糊聚类为M类, 同类 节点可以在同一时隙被调度, 不同类节点必须在不同时

A new hybrid broadcast scheduling scheme for wireless sensor network

Because of the mutual interference and the sharing of wireless links in a wireless sensor network(WSN), conflicts occur when data messages are transmitting between nodes. The broadcast scheduling allocates for each node an independent time interval with a fixed length, and finds an optimal feasible solution of the shortest frame-slot length and the maximal transmission-rate. A two-stage hybrid algorithm is proposed based on the neural network to solve this problem for WSN. In the first stage, a modified sequential vertex coloring algorithm is used to obtain a minimal TDMA(time division multiple access) frame length. In the second stage, a fuzzy Hopfield network is adopted to maximize the channel utilizationratio. Experimental results, obtained from the applications to three benchmark graphs, show that our algorithm can achieve better performance with shorter frame length and higher channel utilization-ratio than other exiting BSP solutions.