Krylov Approximation for Multi-user Detection with Parallel Interference Cancellation in Chip and User Space

Iterative multi-user detection and time-variant channel estimation in a multi-carrier (MC) code division multiple access (CDMA) uplink requires high computational complexity. This is mainly due to the linear minimum mean square error (MMSE) filters for data detection and time-variant channel estimation. We develop an algorithm based on the Krylov subspace method to solve a linear system with low complexity, trading accuracy for efficiency. It has been shown by Dumard et al that this approach enables drastic reduction of computational complexity as well as storage reduction when used for time-variant channel estimation. We compare two scenarios of parallel interference cancellation (PIC), in chip space and in user space. In the case of PIC in chip space, the Krylov subspace method allows parallelization of the computations of the K filters and storage reduction. However, the overall computational complexity remains similar to that of the exact linear MMSE filter. In the case of PIC in user space, it reduces drastically the computational complexity but incurs some loss of performance.