Dual band FSS with fractal elements - Electronics Letters

reverse link transmission model. The single mobile user is assumed to continuously transmit information data at 64kbit/s. The data are divided into a sequence of 5120bit data frames. Frame data are frst 16bit CRC encoded and then turbo encoded with coding rate of 1/3 and generator polynomials of 1, 517, 5/7 in octal notation (the two RSC encoders are the same and have a constraint length of 3 bits (see Fig. la). A multistage interleaver (termed MIL) [4] realises turbo internal interleaving and channel interleaving. At the W-CDMA transmitter, the coded binary data sequence is transformed into a quaternary phase shift keying (QPSK)-modulated symbol sequence after some non-information bearing data have been appended. Four pilot symbols are appended to the resultant 128 ksymboVs QPSK-modulated symbol sequence every 0.625ms (the time interval beginning with four pilot symbols followed by 76 data symbols is called the time slot) and is QPSKspread using a 4.096 Mchip/s complex long pseudo-random spreading sequence (spreading factor 32) to be transmitted over a frequency-selective fading channel [l]. The fading channel is modelled as a frequency selective Rayleigh channel having an ITUspecified vehicular B power delay profile [5]. Two spatially separated antennas and a four-finger (two fmgedantenna) coherent RAKE combiner are assumed at the base station receiver. Channel estimation for coherent RAKE combining is performed by averaging eight received pilot symbols of two consecutive slots. The RAKE combiner output sample sequence is de-interleaved and turbo decoded as described previously. Signal-to-interference plus background noise power ratio (SIR) measurement based fast transmit power control (TPC) described in [6] is assumed; the power up/down step size is 1dB and the power-control rate is 1600Hz. We measured the average bit error rate (BER), average received EJZo per antenna (since the single user is assumed in the simulation, Zo is equal to the background noise spectrum density No), and average number of decoding iterations for various values of target SIR for fast TPC. The measured average BER and the average number of decoding iterations are plotted in Fig. 3 against the average EJNo per antenna for the fading maximum Doppler frequency fD = 5 and 80Hz. The maximum number of decoding iterations is eight (in our simulation, after eight iterations, the improvement in BER performance is saturated). It can be clearly seen from Fig. 3 that the introduction of CRC decoding into the turbo decoding iteration process does not degrade the BER performance at all, while the average number of decoding iterations can be reduced by half at average BER = 1 P and by 75% at average BER = 1W. Similar results were gained under slow fading rates, i.e. fD = 5Hz.