Near optimal scheduling for opportunistic spectrum access over block fading channels in cognitive radio assisted vehicular network

GÜL Ö. M., Kantarci B.

Vehicular Communications, vol.37, 2022 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 37
  • Publication Date: 2022
  • Doi Number: 10.1016/j.vehcom.2022.100500
  • Journal Name: Vehicular Communications
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, INSPEC
  • Keywords: Vehicular network, Cognitive radio, Cross-layer design, Opportunistic spectrum access, Fading, IoT
  • Middle East Technical University Affiliated: Yes


© 2022 Elsevier Inc.With the increasing use of cognitive radio technology in vehicular communications, vehicles will be enabled with cognitive radio in the future. Cognitive radio assisted vehicular networks make cognitive radio enabled vehicles utilize licensed spectrum on highways opportunistically. This work tackles cognitive radio assisted vehicular networks including M primary users (transmitter), M primary receivers, a secondary user (transmitter) with K channels and K secondary receivers. A channel is assigned to each primary user for data transmission whenever they need to use. The secondary user has data backlog. It knows neither the states (good or bad) of M channels, or statistics of channel evolution processes. It selects K of M channels to send data at each time slot. If it senses that a selected channel is in good state, it uses that channel for data packet transmission in same time slot. If not, it chooses another channel at the next time slot. We tackle this problem under average reward criteria for finite and infinite time horizon. This work presents Uniforming Random Ordered Policy (UROP) and shows UROP achieves near-optimal throughput under block fading assumption for generic channel evolution process. Numerical results show that our policy achieves near-optimal throughput (more than 80% of optimal throughput for scarce number of available channels) under average reward criteria whereas MP may achieve less than even 25% of optimal throughput for scarce number of available channels.