Due to the limited power constraint in sensors, dynamic scheduling with data quality management is strongly preferred in long lifetime monitoring applications. But typical techniques treat data management as an isolated process on only selected individual nodes, e.g. the centroid node. In this paper, we propose and evaluate an aggressive data reduction algorithm based on error inference within sensor segments. The architecture integrates three parallel dynamic error control mechanisms to optimize the trade-off between energy saving and data validity. We demonstrate that not only substantial energy savings can be achieved but also that an error bound specified by the application can be guaranteed. Moreover, we have investigate the system performance by using the realistic historical soil temperature data as an experimental context. The experimental results demonstrate that the system error meets the specified error tolerance and produces up to a 50 percent of the energy savings compared to several sensing schemes.

Energy is the most precious resource in wireless sensor networks. To ensure sustainable operations, wireless sensor systems need to harvest energy from environments. The time-varying environmental energy results in the dynamic change of the system's available energy. Therefore, how to dynamically schedule tasks to match the time-varying energy is a challenging problem. In contrast to traditional computing-oriented scheduling methods that focus on reducing computational energy consumption and meeting the tasks' deadlines, we present DEOS, a dynamic energy-oriented scheduling method, which treats energy as a first-class schedulable resource and dynamically schedules tasks based on the tasks' energy consumption and the system's real-time available energy. We extensively evaluate our system in indoor and outdoor settings. Results indicate that DEOS is extremely lightweight (e.g., energy consumption overhead in the worst case is only 0.039%) and effectively schedules tasks to utilize the dynamically available energy.