The problem of distributed and dynamic sensing user selection in cognitive systems is studied in this paper, where channel sensing consumes resources and users behavior is distributed. Since users can obtain the channel state from the fusion center, if there are other users sensing the channel, users may enjoy the results sensed by others rather than sense the channel themselves. Such selfish behavior decreases both network utility and individual rewards. Inspired by the social expectation that no one should always enjoy the fruits of others' labor and that one should provide compensation after obtaining a benefit, we propose a distributed sensing compensation algorithm in this paper. The main concept of this algorithm is that after a channel is accessed successfully, users must sense the channel as a compensation for enjoying others' sensing results. The system state probabilities are obtained using Markov chain analysis. We show that there are always an optimal or near optimal number of users sensing the channel and hence, the near-optimal performance is achieved on average using the proposed algorithm. Additionally, the algorithm achieves good fairness performance with respect to the sensing cost. It is further shown that the proposed algorithm is not only suitable for static scenarios but also adaptable for dynamic scenarios with a changing active user set.
Near Optimal Distributed Cooperative Spectrum Sensing and Access: A Benefit-and-Compensation Approach
2017-09-01
215577 byte
Conference paper
Electronic Resource
English