Goodput-Based Link Resource Adaptation for Reliable Packet Transmissions in BIC-OFDM Cognitive Radio Networks

Abstract

Cognitive radio (CR) stands out as a potential cornerstone to break the spectrum gridlock through enabling the coexistence of licensed (primary) and unlicensed (secondary) users in the same bandwidth. This paper deals with a novel link resource adaptation (LRA) strategy to be applied in CR scenarios for reliable packet transmissions based on bit interleaved coded orthogonal frequency division multiplexing (BIC-OFDM). We first formulate the power allocation (PA) problem constrained by both the available power at the secondary transmitter (ST) and the interference tolerable at the primary receivers, aimed at maximizing the offered layer 3 data rate, i.e., the goodput (GP) metric. Then, we derive the optimal PA strategy resorting to the customary Lagrangian dual decomposition (LDD) technique, which, however, like many other conventional numerical methods, exhibits several drawbacks, such as slow convergence and need for parameter tuning. These restrictions are circumvented through the development of a novel iterative yet simple PA algorithm, referred to as successive set reduction (SSR) approach, whose optimality conditions are analytically demonstrated by resorting to the Quasi Variational Inequality (QVI) framework. Based on this PA algorithm, an adaptive modulation and coding (AMC) scheme at the ST is eventually derived. Simulation results over a realistic scenario corroborate the effectiveness of the proposed SSR-based AMC algorithm, highlighting the GP improvements over non-adaptive LRA techniques, besides a remarkable complexity reduction w.r.t. conventional numerical methods.