A signal detection algorithm based on RLS adaptive AR model in low signal-to-noise ratio condition

Abstract

It is a difficult problem to detect signal in low signal to noise ratio condition. A signal detection algorithm based on AR model is presented. The coefficients of AR model are found using RLS algorithm. The maximum magnitude of coefficients is summed to make the decision variable and compared to threshold. This algorithm has many advantages, such as small requirement of computation, high speed and independent of signal form, etc. Detecting complex sinusoid, QPSK and GMSK signals in white Gaussian noise under low signal to noise ratio condition are simulated. Results show that this algorithm can achieve better performance.