UE autonomous resource selection for D2D communications: Explicit vs. implicit approaches

With the increasing demand for wireless traffic, Device-to-Device (D2D) communications underlaying the cellular network is beneficial to offload traffic and further improve throughput and latency. 3GPP has introduced D2D communications for proximity-based services since Rel-12. The assumption of half-duplex mode and lack of a feedback mechanism make UE autonomous resource selection challenging in terms of collision minimization, which is significant when the eNB is out of function or when UEs move out of coverage. However, the existing design (i.e., Mode 2 communications) fails to address this problem. In this paper, we comprehensively explore UE autonomous resource allocation based on the specified frame structure including Scheduling Assignment (SA) and data transmission. Moreover, we propose the enhanced explicit approach and enhanced implicit approach by utilizing the sensing result to minimize collisions. Simulation results show that the proposed approaches outperform Rel-12 Mode 2 in terms of data collision probability and throughput. The enhanced implicit approach achieves low data collision probability and high throughput by SA admission control to prohibit excess transmissions. The pros and cons among the proposed approaches and Rel-12 Mode 2 are compared. The insights can be further adopted for Vehicle-to-Vehicle (V2V) communications and D2D communications enhancement.