无线多媒体传感器网络实时任务分配算法

Real-Time Task Allocation Algorithm in Wireless Multimedia Sensor Networks

为了满足多媒体信息传输的时延、同步的要求,提出无线多媒体传感器网络(Wireless Multimedia Sensor Networks,WMSNs)在能量受限情况下的实时任务分配算法。根据跟踪目标、传感器节点和汇聚节点的地理位置信息分簇,簇内节点任务分配选择能够满足能量上限的调度长度最小的分配方案。调节任务执行能量上限Emax、合作处理分簇的最大跳数K和执行任务总数Tnum三个参数,随机产生任务图进行仿真实验。结果表明:在性能上,算法满足节点能量受限的要求,在实时性方面有明显改善,与早期的分布式计算架构相比,由仿真图估算可得任务调度长度减少约45%;与带能量限制的任务映射和任务调度算法EcoMapS相比,从仿真图可以看出,任务调度长度也有明显减少。

In order to satisfy the requirement of the network latency of Wireless Multimedia Sensor Networks （WMSNs）, the energy-constrained real-time task allocation algorithm based on coordinates is proposed. The target tracking tasks are allocated to the sensor nodes that have been clustered according to the coordinates of the target, sensors and the sink. The task allocation in the cluster is to choose the scheme which has the minimum task scheduling length. Adjust the maximum energy consumption of the tasks and the maximum hops K in the cluster and the number of total task to do the simulation experiment. Estimated by simulation graph, the results have shown that the performance can satisfy the energy budget, and the task scheduling length reduced 45% while compared with existing Distribute Computation Architecture, and compared with task mapping and scheduling EcoMapS, the scheduling length also reduced a lot.