Impact of Imperfect Spectrum Sensing on Age of Information in Energy Harvesting Cognitive Radios

The metric age of information (AoI) has recently been widely employed to quantify the freshness of the information delivered to the destination. This paper investigates long-term average AoI minimization of an energy harvesting secondary user (SU) in a cognitive radio network setting. Specifically, the paper focuses on the impact of imperfect spectrum sensing on AoI minimization in this setting. The SU makes a decision whether to sense the presence of a primary user, and if it determines the spectrum to be unoccupied, may send out a status update. Sensing and updating both cost energy and the sensing decision may be incorrect due to imperfect spectrum sensing. This setting is formulated as an infinite horizon partially observable Markov decision process (POMDP) to derive the optimal policy that minimizes the long-term average AoI of the SU. The existence of the optimal stationary sensing and update policy is proved and the threshold structure of the policy is shown. Numerical results are presented to demonstrate the SU's AoI performance.