Throughput enhancement in a cognitive radio network using a reinforcement learning method

• Published in: Multimedia tools and applications Vol. 83; no. 1; pp. 1165 - 1187
• Main Authors: Clement, J. Christopher, Sriharipriya, K. C., Prakasam, P., S, Chandra Sekaran D
• Format: Journal Article
• Language: English
• Published: New York Springer US 2024; Springer Nature B.V
• Subjects: Algorithms; Cognitive radio; Communication; Computer Communication Networks; Computer Science; Data Structures and Information Theory; Multimedia; Multimedia Information Systems; Optimization techniques; Probability; Radio networks; Random variables; Sensors; Special Purpose and Application-Based Systems; Spectrum allocation; Teaching methods; Wireless communication systems; Deep deterministic policy gradient; Throughput; Spectrum sensing; Cognitive radio; Reinforcement learning
• ISSN: 1380-7501; 1573-7721
• DOI: 10.1007/s11042-023-15432-8

As the demand for higher data rate is exponentially growing, spectral efficiency improving methods can be adopted in recent day’s wireless communication systems. If the cognitive radio network can forecast the channel to be sensed, instead of sensing all channels sequentially, then reasonable increase in throughput can be achieved. In this research, we forecast not only the channel that can be sensed, but we also predict how long the channel remains usable for secondary users. This process can reduce the sensing time. We use a deep deterministic policy gradient method to optimally select the channel and also the duration for sensing. Doing this way, we can minimise the energy spent on sensing and make the cognitive radio energy efficient. Through simulation, we show that the number of operations invested on sensing is minimised. We also show in our result that the higher throughput is achieved.