A novel scheduling strategy for reconfigurable HW under energy harvesting environment

The scheduling of wireless sensor network (WSN) nodes with energy harvesting is fundamentally different from battery-based nodes. However, reconfigurable hardware has different power consumption modes compared with processor-based implementations. Therefore, an adaptive scheduling strategy is needed for reconfigurable hardware under an energy-harvesting environment. A novel scheduling is proposed to schedule hardware reconfiguration. For WSN implementation, the proposed strategy aims to achieve higher efficiency with less computing requirement. In the proposed strategy, a task window is built first. Then, inside the task window, the execution time for each task is tallied. Only the most used task in the future (MUF) is implemented using reconfigurable hardware. An hardware-reconfigurable WSN node is presented, with related design flow, to demonstrate the proposed scheduling strategy. Four test cases are chosen to evaluate the strategy. The execution of test cases is divided into blocks of 256 bits for each application. Three other strategies (all-hardware, all-software and random) are adopted to implement the evaluation. The results from the experiments show that more than 50% of the energy cost can be saved if the proposed method is implemented.