Spectrum Trading for Energy-Harvesting-Enabled Internet of Things in Harsh Environments.

Energy harvesting refers to energy-supply guaranteed solutions in Internet of Things for functional maintenance, because network nodes can absorb energy from surrounding environments. However, once the environment turns into a harsh condition (e.g., desert, ocean, or battlefield), the design of a robust network model must take into account two intractable factors: 1) the flexibility of users' heterogeneous requirements and 2) the influence of unpredictable emergency in energy harvesting. Consequently, spectrum resource allocation revolved issues become rather challenging. To tackle this problem, we build the central network into a combinatorial auction model where complex interactions between the access point and users can be captured. Meanwhile, our model preserves the user flexibility in time-frequency domain. Based on this model, we demonstrate both the energy supply of access point and the network performance can be guaranteed. Moreover, we propose an approximation algorithm to obtain efficient and effective results, and further provide incentives for users to anticipate a truthful auction. Finally, we modify the proposed model to be practical and yet truthful for pursuing an optimal solution. Extensive simulations validate the effectiveness and robustness of our proposal.