On eigen-based signal combining using the autocorrelation coefficient

In typical signal combining scenarios, the combining weights are estimated using the criterion of maximum average signal-to-noise ratio (SNR) or maximum combined output power (COP). Eigen-based algorithms are very important and popular in signal combining. The conventional SNR EIGEN or COP EIGEN may not necessarily be effective in terms of performance or system complexity. The main contribution of this study is the introduction of the combined signal autocorrelation coefficient as a newer objective function to signal combining. The corresponding eigen-based combining algorithm AC EIGEN and its modified algorithm MAC EIGEN are also derived. Proposed algorithms have the same simple system structure as the COP EIGEN, which can successfully avoid estimating the noise correlation matrix. Simulation results indicate that the AC EIGEN and the MAC EIGEN have good combining performance for signals with white Gaussian noise when the SNR of the signals is low. Considering the system complexity of the SNR EIGEN, and the COP EIGEN being biased for non-uniform noise variance signals, the proposed algorithms are attractive.