Device-to-device resource scheduling by distributed interference coordination

Device-to-device (D2D) communication overlaying a cellular network has been proved to augment the flexibility and enable new applications (e.g., content distribution) in cellular networks. In this context, it is crucial to design a radio resource management system that allocates the time-frequency (TF) resources to the D2D links in a distributed way, i.e. without any coordination by the cellular network, so as to guarantee the quality of service (QoS), particularly in heterogeneous traffic conditions. Interference-aware resource allocation has the capability to adapt the resource management to a context where multiple D2D links coexist in the same spectrum. In this paper, a distributed scheduling approach is proposed where each D2D link reacts to the locally sensed interference by self-adapting its own TF allocation. Each node autonomously trades the QoS request in term of packet service with the resource availability by inflating/deflating the spectrum allocation based on the sensed interference level. The change of the interference pattern perceived in turn by other D2D links serves as interlink signaling of the need/release of TF resources. Each node optimizes the allocation by iterated local adjustments, till an equilibrium with other D2D links is reached. The paper shows that the proposed scheduling algorithm is able to maximize the total throughput in a fully distributed way, arranging efficiently the radio resources over the TF domain so as to satisfy the QoS requirements for each node.