A two-layer coalitional game among rational cognitive radio users

Abstract

In cognitive radio (CR) networks, secondary users (SUs) sense the spectrum to identify and possibly transmit over temporarily unoccupied channels that are licensed to primary users (PUs). However, when the received PU signal is weak, spectrum sensing by individual SUs becomes unreliable. To improve sensing accuracy, SUs can form disjoint coalitions and cooperate to discover idle time slots. These spectrum opportunities are then shared among the coalition members in a coordinated manner. It is proposed to decouple the coalition formation and the access (payoff) allocation problems by modeling these processes as a two-layer coalitional game. This game fosters cooperation by providing each SU with the access opportunities it deserves. Numerical results demonstrate that the proposed game outperforms previously investigated collaborative sensing and access approaches in terms of energy efficiency, throughput, and fairness.

A Sensing Contribution-Based Two-Layer Game for Channel Selection and Spectrum Access in Cognitive Radio Ad-hoc Networks

“…Thus, large values of |S| are penalized, balancing SU competition on all channels. Finally, our extensive simulation results using (12)(13)(14)(15) show that the AND-rule is more suitable when cooperative SUs have heterogeneous sensing capabilities (i.e., different PU-to-SU SNRs) while the OR-rule provides better performance in homogeneous scenarios. Intuitively, if there is at least one SU m ∈ η with favorable PU-to-SU SNR λ mn , and thus with small FA probability P n FA,AND (m), the resulting P n FA,AND (η) is very small since P n FA,AND (η) ≤ min m∈η P n FA,AND (m) in (12).…”

“…For SU m on channel n [cf. (12) and (14)], this rate is determined by λ mn and SU population size |C n t (1)|, where t is the time slot index. Each SU m on channel n maintains three values of its FA rate, corresponding to the current and potential SU populations sizes |C n t (1)| ± i, i = 0, 1, and this FA rate information is exchanged over the control channel as needed.…”

“…While we have first proposed the two-layer game in [14], that paper did not contain several key proofs, practical validation, and complexity analysis of the coalitional game and did not address sensing data fusion rules for heterogeneous environments.…”

“…The proof of Proposition 5 is based on Propositions 2-4 and follows from the superadditivity property [16] for the 0/X model and the weak efficiency or the nonpositive externality property [23], [24] for the 1/X model [14] 10 . Proposition 5 implies that all SUs sensing channel n cooperate, and MAC is not utilized.…”

“…≥ 10) requires less than 10 samples for good performance (See, e.g.,[18],[22]). 9 Proofs of Propositions 2 and 5-8 can be found in[14].…”