Abstract:
A method and device for handling write access conflicts in interleaving, in particular for high-throughput turbo decoding for wireless communication systems; the device comprises N interleaving buffers (CLk) that are respectively connected to N producers (PRk), an LLR distributor means and N single port target memories (TMk). At any time step, each interleaving buffer receives m LLR inputs from the producers and has to write up to M of these into a register bank (RBk), which comprises W registers. M denotes the maximum number of concurrent write operations supported per time step and W denotes the maximum buffer size. M and W are design parameters and are chosen for the standard case and not for the worst case. m-M producers have to be stalled whenever m is larger than M and m producers have to be stalled whenever a buffer overflow occurs (more than W LLR values). Finally, at any time step one LLR value is fetched from the register bank and written to the SRAM interleaving memory.
Abstract:
In a first step, slot synchronization is obtained by setting in correlation (210, 220) the received signal (r) with a primary sequence (SG), which represents the primary channel (PSC), and by storing said received signal. During a second step, the said correlator (210, 220) is re-used for correlating the received signal (r) with a secondary sequence (SSC) corresponding to the secondary synchronization codes. The correlator (210) is preferably structured in the form of a first filter (210) and of a second filter (220) set in series, which receive a first secondary sequence (SG1) and a second secondary sequence (SG2), typically consisting of Golay sequences. Proposed herein are architectures of a parallel and serial type, as well as architectures designed for re-using further circuit parts. A preferential application is in mobile communication systems based upon standards such as UMTS, CDMA2000, IS95, or WBCDMA.