A Non-Data-Aided Decision-Directed Feedforward Symbol Timing Estimation – Implementation Issues

W.K. Wong (PRC)


Symbol timing estimation, non-data-aided timing estimation, feedforward timing estimation modelling and simulation. I


: Various implementation issues for decision-directed timing estimation (DD) are addressed in this paper. These include the theoretical evaluation of DD's signal, the sampling rate, the number of the symbols for each timing estimation, the normalization process, and the quantization effect. In this paper, four samples per symbol are used. Based on a look-up table, the computation is simple but the performance is excellent at small roll-off factors. Computer simulation using QPSK signal shows that the degradation of timing variance due to the quantization effect in DD is far better than that for O&M and LEE, and is marginally better than that of ML-LOGN when 4 bits resolution is used for the sampled data and 15% excess bandwidth is used for the raised cosine filter.

Important Links:

Go Back