An Intelligence-Based Recurrent Learning Scheme for Optimal Channel Allocation and Selection in Device-to-Device Communications

Abstract

Decentralized mobile user communications rely on sensing and signal processing that are aided by fusion centers. Device-to-device (D2D) communications form the backend network for facilitating mobile user communication through the appropriate signal sensing and channel selection. The sensing and processing of tightly coupled channels ensure seamless uninterrupted communications with fewer outages. An imbalance in channel allocation and selection increases the gap between communication networks and signal processing systems. Unattended channel allocation and selection result in delayed communications and additional power exploitation with less reliability. This paper introduces an intelligence-based recurrent learning (IRL) scheme for optimal channel allocation and selection for mobile users’ D2D communication. The objective of this paper is to select a delay-controlled channel satisfying both the data rate and power control requirements in the channel allocation. The allocated channel is analyzed through a responsive linear system transformation for its power, data rate, and time constraints in a recurrent manner. The intelligent learning technique evaluates the consistency of the channel based on a recurrent analysis. The outcome of the analysis is the selection of an optimal channel from the allocated channels that satisfies the objective and channel policies. Synchronized channel allocation achieves power-controlled communications in a cooperative manner under controlled interference. The proposed IRL minimizes D2D communication delay, transmits the power requirement and outage, and improves throughput with better reliability.