Correlation-Index Blind Classification for MFSK with Unreconstructed Compressive Samplings

—To improve the speed of traditional signal sampling and processing, which is based on the Nyqusit-sampling theorem, a new theorem called Compressive Sensing (CS) is presented. Most work in CS has focused on reconstructing the high-bandwidth signals from nonuniform low-rate samples. However, the signal reconstruction process of CS needs large amount of computation. In this paper, under the frame of CS, a method called Correlation-Index Blind Classification (CIBC) using the Correlation-Index to realize MFSK signal classification is presented, which exploits the compressive measurements without reconstructing the original signal or knowing any parameter. CIBC processes the compressive measurements directly, avoiding the high computation of reconstruction. What’s more, compared with the traditional Peak-Index, the promotion of computing speed is significant, and the performance decreases very little. The simulation results show that CIBC has good performance and low computation complexity. It’s an effective method to improve the real-time performance of communication system.