Optimal resource allocation for sensing-based spectrum sharing D2D networks

To improve the spectral efficiency of a cellular network, sensing based spectrum sharing device-to-device (sensing-based D2D) technique is introduced in this paper. To be specific, efficient algorithms are introduced to find the optimal sensing time and power allocation for single-subband sensing-based D2D systems under cellular link’s rate loss constraint. For multi-subband systems, D2D users may have limited sensing capability where only a part of the subbands can be sensed. Accordingly, a partially observable Markov decision process (POMDP) model is introduced and an efficient algorithm is introduced to address the resource allocation problem.Simulation results suggest that sensing-based D2D outperforms overlay D2D when the transmitter of D2D link is far away from the receiver of cellular link in the single-subband single-subcarrier case. In the single-subband multi-subcarrier case, sensing-based D2D can achieve 15.2% increase in spectral-efficiency over overlay D2D. In the multi-subband case, when D2D users have limited sensing capability, our proposed Separate Principle method can further improve D2D link’s spectral-efficiency by 5%∼15%5%∼15% compared to the Greedy algorithm.