Device to device (D2D) Communication is essential in emerging networks, such as Internet of things (IoT). Cognitive radio (CR)-enabled device is a promising technology to address D2D Communications. However, existing work on CR networks suffers from impractical protocol designs which lead CR users to long spectrum access delay (SAD) with high energy consumption (EC). In this research, a practical adaptive framework for fast and energy-efficient spectrum access for D2D Communications is explored. The design challenges are addressed from two perspectives. First, new protocols on the medium access control (MAC) layer are established in order to deal with the unique issues in the MAC, such as the rendezvous failure due to channel status change, the handshake failure due to asynchronous time slots, the unaware deadlock due to the handshake message collision, the transmission failure due to the restless handoff, and the network congestion due to the packet congestion in each CR device. Another reason causing the long SAD with high EC is the long restless operation of the rendezvous process itself, since existing channel hopping schemes for rendezvous mainly focus on the hopping sequence design without considering the node conditions and communication environments. Therefore, cross-layer rendezvous design jointly considering channel selection, channel hopping, and transmit power control is studied.The proposed practical adaptive framework in this research is endowed with the learning and mining abilities, which makes unlicensed users truly cognitive. Therefore, this research will provide important insights on the designs of future D2D Communications.
