Premium
A joint sensing and transmission power control policy for RF energy harvesting cognitive radio networks
Author(s) -
Yan Feiyu,
Zhao Jihong,
Qu Hua,
Xu Xiguang
Publication year - 2018
Publication title -
international journal of communication systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.344
H-Index - 49
eISSN - 1099-1131
pISSN - 1074-5351
DOI - 10.1002/dac.3715
Subject(s) - cognitive radio , computer science , retransmission , channel (broadcasting) , partially observable markov decision process , markov decision process , computer network , transmission (telecommunications) , fading , power control , transmitter power output , network packet , energy harvesting , data transmission , markov chain , energy (signal processing) , markov process , wireless , telecommunications , transmitter , markov model , power (physics) , mathematics , statistics , physics , quantum mechanics , machine learning
Summary We consider a radio frequency energy harvesting cognitive radio network in which a secondary user (SU) can opportunistically access channel to transmit packets or harvest radio frequency energy when the channel is idle or occupied by a primary user. The channel occupancy state and the channel fading state are both modeled as finite state Markov chains. At the beginning of each time slot, the SU should determine whether to harvest energy for future use or sense the primary channel to acquire the current channel occupancy state. It then needs to select an appropriate transmission power to execute the packet transmission or harvest energy if the channel is detected to be idle or busy, respectively. This sequential decision‐making, done to maximize the SU's expected throughput, requires to design a joint spectrum sensing and transmission power control policy based on the amount of stored energy, the retransmission index, and the belief on the channel state. We formulate this as a partially observable Markov decision process and use a computationally tractable point‐based value iteration algorithm. Section 5 illustrate the significant outperformance of the proposed optimal policy compared with the optimal fixed‐power policy and the greedy fixed‐power policy.