Adaptive task scheduling for providing event miss-ratio statistical assurances in wireless sensor networks

This paper addresses event miss-ratio statistical assurances for data gathering in cluster based dense wireless sensor networks. In line with recent commercial off-the-shelf sensor nodes, this paper considers sensor nodes with multiple sensing units and power modes of operation. Such sensor nodes are usually unable to process the data generated by multiple sensing units simultaneously, thereby resulting in event misses. We present an adaptive task scheduling protocol, called ASPEN, for providing event miss-ratio statistical assurances in wireless sensor networks. ASPEN consists of two coarse-grain task scheduling algorithms at the cluster-level and one fine-grain task scheduling algorithm at the node level. The first coarse-grain scheduling algorithm eliminates misses during event processing by assigning each sensor node the minimum number of event types to observe. The other coarse-grain scheduling algorithm enables accurate transfer of data from sensor node to the cluster-head for each event type while ensuring that less redundant data are transmitted. The proposed fine-grain scheduling algorithm determines the order of data processing for multiple sensing units, chooses appropriate power modes to avoid event misses, and takes the effect of environmental factors such as temperature on the sensing ability of the nodes into account. Simulation results show that, with sufficient node redundancy, ASPEN can provide statistical assurances for event miss-ratio while reducing the amount of event data transmitted up to 65%