Code-aided ML joint delay estimation and frame synchronization Henk Wymeersch and Marc Moeneclaey Digital Communications Research Group Dept. of Telecommunications and Information Processing Ghent University, Sint-Pietersnieuwstraat 41, 9000 GENT, BELGIUM E-mail: {hwymeers,mm}@telin.ugent.be Abstract We present a novel maximum-likelihood (ML) algorithm for joint delay estimation and frame synchronization. The algorithm operates on coded signals, and exploits the code properties by accepting soft information from the MAP de- coder. Issues of convergence are addressed and we show how computational complexity may be reduced without any performance degradation. Simulation results are presented for convolutional and turbo codes, and are compared to per- formance results of conventional algorithms both in terms of mean square estimation error (MSEE) and BER. We show that code-aided delay estimation always improves the MSEE, but not necessarily the BER. On the other hand, code-aided frame synchronization is mandatory, in order to avoid either significant BER degradations or the need for very long pilot sequences. Keywords: turbo synchronization, delay estimation, frame synchronization, EM algorithm Contents 1 Introduction 2 2 System Description 2 3 ML estimation through the EM algorithm 3 4 Code-aided DE and FS 3 5 Conventional DE and FS 4 6 Convergence properties 5 7 Performance results 5 7.1 Computational complexity and additional convergence issues ........................ 6 7.2 Delay estimation ............................................... 6 7.3 Frame synchronization ............................................ 6 7.4 Joint delay estimation and frame synchronization .............................. 7 8 Conclusion and Remarks 7 1