Feature Extraction of the Radio Signals Using Attribute Reduction of Formal Concepts

Radio signal recognition is an important aspect of radio signal processing. Feature extraction of radio signals is a key of radio signal recognition. In this paper, the formal context is proposed to describe the relation between signals and their features. Therefore, the theory and method of attribute reduction of formal concept lattice can be utilized to extract useful features. Attribute reduction of formal concept lattice based on similarity relation of attributes is discussed. The sufficient and necessary conditions, which are used to justify whether an attribute is the core, relative necessary or absolutely unnecessary, are proposed. All of these can be used to extract the core features of the radio signals. The experimental results show that we can easily and effectively extract useful features of radio signals by the proposed method in this paper.