A cross-uncorrelator-initializer for the super-exponential algorithms in multi-user environment

The super-exponential algorithms (SEAs) have potential to equalize a communication channel and to mitigate interference in a multi-user environment, and possess very fast (exponential) convergence rate. By considering the basic principle behind the development of SEA, we propose a novel initialization scheme based upon the null space of a cross-correlation matrix to prevent repeated retrieval of identical sources. A sequential signal-canceler based cascaded-equalizer structure is therefore not required. Simulation results support the proposed method.