Distributed Cognitive MAC for Energy-Constrained Opportunistic Spectrum Access

We address the design of distributed cognitive medium access control (MAC) protocols for opportunistic spectrum access (OSA) under an energy constraint on the secondary users. The objective is to maximize the expected number of information bits that can be delivered by a secondary user during its battery lifetime without causing interference to primary users. By absorbing the residual energy level of the secondary user into the state space, we formulate the energy-constrained OSA problem as an unconstrained partially observable Markov decision process (POMDP) and obtain the optimal spectrum sensing and access policy. We analyze and reduce the computational complexity of the optimal policy. We also propose a suboptimal solution to energy-constrained OSA, whose computational complexity can be systematically traded off with its performance. Numerical examples are provided to study the impact of spectrum occupancy dynamics, channel fading statistics, and energy consumption characteristics of the secondary user on the optimal sensing and access decisions