Practical scheduling for stochastic event capture in energy harvesting sensor networks

Existing scheduling schemes for stochastic event capture with rechargeable sensors either adopt simplified assumptions on event's properties or provide no performance guarantee. Considering the stochasticity of event staying time and event capture utility, we investigate the sensor scheduling problem aiming to maximise the overall quality of monitoring (QoM) in event capture application of energy harvesting sensor networks. We first provide a paradigm to calculate the QoM of a point of interests (PoI) and formulate the scheduling problem as an optimisation problem. Although we find that this problem is NP-complete, we prove that the problem can be cast as maximisation of a submodular function subject to matroid constraints. Accordingly, we can design centralised and distributed algorithms, each of which achieves a factor of 1=2 of the optimum. We evaluate the performance of our solution through simulations, and simulation results show that our scheme outperforms former works.