D. Emir, B. Abdellatif, and B. Ammar (Tunisia)
PRNN, W-CDMA, Interference estimation.
In this paper, we propose an non-linear adaptive structure based on Pipelined Recurrent Neural Network (PRNN) for multi-user receiver for the downlink in W-CDMA . The PRNN is so called because of its recurrent and its modular processing. The Recurrent processing is used in order to more accurately estimate the signal caused by the Multi Access Interference (MAI), the Intersymbol Interference (ISI) and the channel . The modularity of the network helps us to contain the computational complexity. The regularity of the PRNN simplify the design of a hardware implementation. Simulations will be done to choose the PRNN dimension and to compare the performances of the proposed receiver with MMSE receivers.
Important Links:
Go Back
