Abstract: A novel dynamic spectrum access technology based on multi-features model clustering is presented to improve transmission quality for different users in the cognitive network. After the multi-features models of the channels are extracted, these channels suitable for the cognitive users are clustered into different spectrum poolings using a multi-objective clustering algorithm. Those channels in the white pooling are prior to be accessed for cognitive users. Multiple feature combinations and dynamic spectrum simulations are carried out to verify the efficiency of the proposed method. Analysis and simulation results show that the channels can be classified efficiently according to their performance. The access efficiency can be improved by the proposed method.

Key words: communication, cognitive radio, multi-features model, spectrum pooling, dynamic spectrum access