Suppression of non-Gaussian narrowband interference in CDMA systems via an interior point based method

In this paper, we consider the problem of narrowband interference suppression in CDMA systems. The unknown narrowband interference is modeled as a non-Gaussian autoregressive process and the received spread spectrum signal is modeled as a finite-state Markov chain. We propose an iterative, off-line scheme based on the coordinate descent method. We estimate the spread spectrum sequence using the Viterbi algorithm, and then estimate the unknown narrowband interference using an interior point based scheme that minimizes the l₁ norm of the residuals. The procedure is repeated until convergence