Use this URL to cite or link to this record in EThOS:
Title: Real-time frame synchronisation in OFDM modems using artificial neural networks
Author: Al-Meer, M. H.
Awarding Body: University of Wales Swansea
Current Institution: Swansea University
Date of Award: 2000
Availability of Full Text:
Access from EThOS:
The problem of frame synchronization in ODFM modems is a critical matter since it determines whether the data received are in the wrong or the right format. OFDM, like any other digital modulation technique, produces its modulated data in a structure of frames. And since frames are transmitted down the channel, it is important to identify the first data in the frame. Therefore the need for timing synchronization occurs in OFDM modem systems. Timing or symbol synchronization has always been a topic of research for computer-communication researchers, who have proposed many solutions. These solutions have different abilities to withstand noise and multi-path deformations encountered in the wireless channel. At the present time, three categories of solution for frame synchronisation exist. The first one depends on sensing the presence of a signal on the channel using a threshold detector. The second solution depends on transmitting a known synchronisation symbol and the receiver correlates the incoming signal to a copy of the synchronisation symbol. The third solution is similar to the second, but the receiver measures the correlation for the OFDM signal to itself. Using the cyclic extension in the OFDM symbol, the detector can sense the presence of the synchronisation pattern. This thesis proposes an efficient frame-timing synchroniser for the OFDM system. The proposed technique is robust against noise and multi-path propagation. The technique is based on an Artificial Neural Network (ANN) as a frame synchroniser. The ANN structure of multi-layer neural networks can indicate the presence of the synchronisation pattern (Wobbulation signal pattern) efficiently, as the results show. A simulation model has been constructed to ease the design of the ANN frame detector. In addition a real-time implementation for the ANN model has been constructed using the TMS32050 DSP processor. The simulation and implementation give results in close agreement, which prove the robustness of the method against noise and multi-path effects.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available