Key technologies of Wideband Code Division Multiple Access (WCDMA)

The key technologies for WCDMA industrialization include RF and baseband processing technologies, specifically including RF and IF digital processing, key technologies such as RAKE receivers, channel coding and decoding, power control, and enhancement technologies such as multi-user detection and smart antennas.

RF and IF

The RF part is a traditional analog structure to realize RF and IF signal conversion. RF uplink channel part mainly includes automatic gain control (RF part is a traditional analog structure, realizing RF and IF signal conversion. The RF uplink channel section mainly includes automatic gain control (RFAGC), receive filters (Rx filters) and downconverters. The RF downlink channel part mainly includes the secondary upconverter, broadband linear amplifier and RF transmitter filter. The IF section consists of the uplink demixing filter, downconverter, ADC and downlink IF smoothing filter, upconverter and DAC.Unlike GSM signals and first generation signals, WCDMA signals are broadband signals with a bandwidth of up to 5 MHz. The linearity and efficiency of RF amplifiers for wideband signals is a common contradiction.

RAKE RECEIVER

The RAKE receiver is a classical diversity receiver designed for CDMA systems, based on the theory that multipath signals can be practically regarded as uncorrelated when the propagation delay is more than one codec period.

The correlator with DLL is a phase-locked loop with late and early gates. It consists of two correlators (early and late) that are ±1/2 (or 1/4) code slice apart from the demodulation correlator, respectively. The correlation result of the early and late gates is subtracted and can be used to adjust the code phase. The performance of the delay loop depends on the loop bandwidth.

The delay estimation serves to obtain the distribution of signal energy at different time-delayed locations by means of matched filters, to identify multipath locations with larger energy, and to assign their time quantities to different receive paths of the RAKE receiver. The matched filter can be measured with an accuracy of 1/4-1/2 code slice, while the interval between the different receive paths of the RAKE receiver is one code slice. In practical implementation, if the delay estimation is updated very quickly (e.g., tens of ms at a time), the phase-locked loop of the sooner-or-later gate can be dispensed with.

Due to the effect of fast fading and noise in the channel, the phase of the actual received paths varies greatly from the phase of the original transmit signal, so the phase has to be rotated according to the result of channel estimation before merging.The channel estimation in the actual CDMA system is accomplished based on the guide symbols carried in the transmit signal. Depending on whether or not the transmit signal carries a continuous guide frequency, phase prediction methods based on continuous guide frequency and phase prediction methods based on judgment feedback techniques can be used, respectively.

The search and demodulation of multipath is performed for each user in the system, and the code slice rate of WCDMA is very high, and the processing of its baseband hardware is very large, which has some difficulties in practical realization.

Channel coding and decoding

Channel coding and decoding mainly reduces the signal propagation power and solves the inevitable fading problem of the signal in the wireless propagation environment. Coding and decoding techniques combined with the use of interleaving techniques can improve the BER performance, compared with the no-coding case, the traditional convolutional code can increase the BER by two orders of magnitude to 10-3~10-4, and Turbo code can further increase the BER to 10-6. WCDMA candidate channel coding and decoding techniques originally included Reed-Solomon and Turbo code, Turbo code because of the coding and decoding, the codec can be used to reduce the signal power and solve the inevitable fading of the signal in the wireless propagation environment. Turbo codes were finally adopted as the data codec for 3G because the codec performance is close to the Shannon limit. Convolutional codes are mainly used for voice and signaling at low data rates. turbo code consists of two or more basic encoders cascaded in parallel by one or more interleavers, as shown in Figure 3:

Turbo code is based on the algorithmic and structural modifications of the traditional cascade code, and the introduction of the inner interleaver makes it possible to eliminate the positive feedback of the iterative decoding. the iterative decoding algorithm of Turbo includes SOVA (SOVA), which is an algorithm for the iterative decoding of a signaling code. decoding algorithms include SOVA (Soft Output Viterbi Algorithm), MAP (Maximum A Posteriori Probability Algorithm) and so on. The iterative decoder of the MAP algorithm achieves greater coding gain since the performance improvement of each iteration of the MAP algorithm is superior to that of the Viterbi algorithm. The actual implementation of the MAP algorithm is the Log-MAP algorithm, which reduces the amount of computation by placing the MAP algorithm in the logarithmic domain for computation.

The difficulty in realizing the Turbo decoding algorithm lies in the decoding rate and the corresponding number of iterations at high data speeds. existing DSPs have built-in the basic algorithms required for the decoder, which allows Turbo decoding to be implemented directly on DSP chips without the need for ASICs.