摘要
为了解决目标跟踪视频传感器网络(VSNs)实时性、有限能量和处理能力的问题,提出基于位置的任务分配算法.采用有向无环图来描述周期的目标跟踪任务,根据跟踪目标、传感器节点和汇聚节点的位置坐标形成合作处理簇,优化簇内节点的任务分配.调节合作处理簇的最大跳数和任务完成时限2个参数进行仿真实验,结果表明,算法能够根据参数动态地组织传感器节点并行计算,提高网络处理能力.在性能上不仅满足任务实时性要求,而且比分布式计算架构节约28%的能量消耗.
A task allocation algorithm based on coordinates was proposed to solve the problems of real-time,limited energy and processing power in target tracking video sensor networks (VSNs).The cyclic target tracking tasks are described by directed acyclic graph and allocated to the sensor nodes that have been clustered according to the coordinates of the target,sensors and the sink.Simulation was performed by adjusting the maximum hops in the cluster and the maximum delay of task execution.The results showed that the algorithm can dynamically organize the sensors to compute in a parallel model according to the parameters.The performance can not only satisfy the real-time requirement,but also reduce the energy by 28% compared with the distribute computation architecture.
出处
《浙江大学学报(工学版)》
EI
CAS
CSCD
北大核心
2010年第4期670-674,共5页
Journal of Zhejiang University:Engineering Science
关键词
视频传感器网络
协同网内处理
任务分配
目标跟踪
video sensor networks
collaborative in-network processing
task allocation
target tracking