1st International ICST Workshop on Cognitive Wireless Networks

Research Article

Resource Allocation for Cognitive Radio: A nonlinear programming approach

  • @INPROCEEDINGS{10.1145/1577382.1577385,
        author={Pin-Hsun Lin and Tong-Hua Hsieh and Hsuan-Jung Su},
        title={Resource Allocation for Cognitive Radio: A nonlinear programming approach},
        proceedings={1st International ICST Workshop on Cognitive Wireless Networks},
        publisher={ACM},
        proceedings_a={CWNETS},
        year={2007},
        month={8},
        keywords={},
        doi={10.1145/1577382.1577385}
    }
    
  • Pin-Hsun Lin
    Tong-Hua Hsieh
    Hsuan-Jung Su
    Year: 2007
    Resource Allocation for Cognitive Radio: A nonlinear programming approach
    CWNETS
    ACM
    DOI: 10.1145/1577382.1577385
Pin-Hsun Lin1,*, Tong-Hua Hsieh1,*, Hsuan-Jung Su1,*
  • 1: Department of Electrical Engineering, National Taiwan University, Taipei, Taiwan, 10617
*Contact email: f89921145@cc.ee.ntu.edu, b91901113@cc.ee.ntu.edu, hjsu@cc.ee.ntu.edu

Abstract

The design goal of the cognitive radio (CR) is to enhance the spectrum use under different environments. To achieve higher spectrum use with negligible interference to legacy users, we consider the joint resource allocation, i.e. waveform design and power allocation, for CR users. To endow the CR systems with high flexibility to work under different environments, we adopt the use of optimization framework in this paper. The optimization target is the sum rate of the spectrum where the CR and legacy users coexist. The quality of service of legacy users is the main constraint and waveforms and powers are the variables. Different constraints from different environments can be easily taken into account by this framework. However, this optimization problem is not convex. To take the advantage of convex optimization, we develop an algorithm to transform it into a convex one with a small approximation error. We also propose an iterative method to reduce the complexity of solving the convex problem. The flexibility and efficiency of the proposed method makes it as a strong candidate for the implementation of CR systems.