Dynamic approaches for spectrum control in cognitive radio networks

Cognitive Radio is an innovative wireless engineering paradigm which promises to mitigate spectrum scarcity problem. It provides wireless systems with capabilities of manipulating their transmission parameters to achieve the highest possible spectrum utilization. This is done by allowing secondary users (unlicensed users) to access licensed spectrum if the primary user (licensed user) is not using it. One necessary condition has to be satisfied. Secondary users have to guarantee that no harmful interference is caused to the primary user. Three important challenges arise when developing such cognitive radio systems. First, electromagnetic spectrum used for wireless communication is very large. Looking for unused channels (sensing channels to figure out their status) is a very expensive operation. Any effective cognitive system has to implement efficient sensing order scheduler to reduce spectrum probing cost. The second and the third challenges arise from the coexistence of heterogeneous secondary users. These users have to share the unused spectrum fairly. Also, they must access the medium in an organized fashion to minimize collisions overhead. This research investigates these three challenges extensively. It models the first problem as Multi-Arm Bandit problem and Regret concept is used to evaluate sensing order scheduling performance. Two artificial intelligence techniques are utilized which are Fuzzy Inference and Bayesian Inference. To optimize Fuzzy Inference scheduler, Genetic Algorithm is used. Compared to fixed sensing order,this proposed technique managed to achieve an outstanding performance in term of channel utilization. Another two solutions are proposed to solve problem of co-existing heterogeneous secondary users. The first solution can be used in centralized fashion where a central entity exists which decides transmission power for all secondary users. This solution has one objective which is to minimize the time required by secondary users to clear their queues. Interior-Point Method is used to find the best spectrum sharing and medium access policy to achieve this objective. The second solution assumes the autonomy of secondary users where the decision to update transmiss on power is distributed among users. Dynamical system approach is used to model system behavior and a forecasting engine based on Deep Neural Network is proposed. This engine gives secondary users the ability to acquire useful knowledge from surrounding wireless environment. As a result, better transmission power allocation is achieved. Evaluation experiments have confirmed that adopting Deep Neural Network can greatly improve the performance.

Text FK 2015 80 D.pdf Download (1MB)

Actions (login required)

View Item