Energy-efficient task scheduling for wireless sensor nodes with multiple sensing units

It is critical to design efficient power management and scheduling techniques for wireless sensor nodes that operate with limited battery power. The design of these techniques becomes more challenging especially when each sensor node has multiple sensing units and the occurrences of their events of interest are not known in advance. This paper proposes an event-driven task scheduling algorithm, called MSUS (multiple sensing unit scheduling). MSUS determines the best power state based on the timing and priority requirements, the event-misses, and the existing and the predicted future tasks for all the sensing units of a sensor node with multiple sensing units. Simulation results show that MSUS leads energy savings up to 49% and reduces event misses by 56% when compared to the greedy technique that puts the sensor node components into the lowest power consumption level whenever there is no available task to process.