Utilizing residual sensing slots to enhance energy efficiency of opportunistic cognitive radio networks

Abstract

In Broadband Communications (BBC), terms like “Green Communication (GC)”, “Energy Efficiency (EE)”, and “Opportunistic Cognitive Radio (OCR)” are pursuing methods for saving energy consumption while accomplishing communications. Good implementations of these terms/concepts contribute in reducing pollution and preserve nature resources. Forthcoming mobile generations will face many tradeoffs between EE considerations versus system performance parameters especially throughput. In a multi-channel OCR system, the Secondary User (SU) senses some selected candidate primary channels based on a “belief” concept. This belief state represents the occupancy/vacancy state of the primary channel and is deduced via a Partial Observation Markov Decision Process (POMDP). Within the OCR transmission frame, a fixed time-period at the beginning of this frame is dedicated for sensing multi primary channels. The SU performs the sensing needed to explore the state of occupancy/vacancy/fading of primary channels -within the sensing period- by dedicating a sensing slot for sensing each primary channel, then, SU updates its channel state belief accordingly. SU may not find it necessary to continue sensing more channels -using more sensing slots- in case of “low belief”, “bad CSI”, “highly believed to be occupied”, or even “sleep” if no candidate channels exist. This will lead to residual (not used) sensing slots which will be a wasted time in the frame time interval. This paper proposes an enhanced algorithm (to a previous work Feng and Gan, 2015) where these unused sensing slots are utilized to enhance the system EE. This is done by manipulating extended variable time induced from using these residual slots to extend the available transmission time. This induced variable time will increase throughput while keeping sensing energy. The metric used in this work is the Normalized Energy Efficiency (NEE) which is in consistence with many research directions in this field. In Feng and Gan (2015), the obtained NEE results reached ∼93% after about 20 transmitted frames, as well as, when 20 primary channels exist in the system. This work proposed an enhanced optimum and approximate algorithms that obtain as high as ∼135% of NEE after about 20 transmitted frames, they also obtained ∼228% using the optimum algorithm, or, ∼138% using the approximate algorithm when 20 primary channels are available. Keywords: Green cellular, Energy efficiency, Cognitive radio, Partially observable Markov decision process, Heterogeneous wireless networks