Improve the Sensing Matrix Model for Random Demodulation in the Case of Mixer With Non-Ideal Characteristics

Abstract

The random demodulation structure successfully applies compressed sensing technology to analog signal sampling, and is expected to replace the Nyquist sampling frame as a new data acquisition scheme. In practical applications, the nonideal characteristics of the mixer will affect the sensing matrix in the RD sampling structure, especially when processing radio frequency signals. This paper analyzes this problem in detail, and proposes a method to modify the sensing matrix to improve the model. Simulation experiments show that the more complete sensing matrix model proposed in this paper can correct the influence caused by the non-ideal characteristics of the mixer.