Iterative detection of turbo-coded offset QPSK in the presence of frequency and clock offsets and AWGN

Abstract

The key contribution of this work is to develop transmitter and receiver algorithms in discrete-time for turbocoded offset QPSK signals. The proposed synchronization and detection techniques perform effectively at an SNR per bit close to 1.5 dB, in the presence of a frequency offset as large as 30 % of the symbol-rate and a clock offset of 25 ppm (parts per million). Due to the use of up-sampling and matched filtering and a feedforward approach, the acquisition time for clock recovery is just equal to the length of the preamble. The carrier recovery algorithm does not exhibit any phase ambiguity, alleviating the need for differentially encoding the data at the transmitter. The proposed techniques are well suited for discrete-time implementation. Index TermsOffset QPSK (quadrature phase shift keying), frequency offset, clock offset, synchronization, matched filtering, additive white Gaussian noise (AWGN). I. INTRODUCTIONGeosynchronous satellites provide line-of-sight communications with the ground stations. Such communication links offer distortionless transmission, with the only impairment being AWGN. Whereas transmit power is not so much of an issue at the ground station, it is a precious commodity on-board the satellite. With the growing demand for satellite broadcast services, it has become necessary for the end users to receive signals directly from the satellite. This calls for a vastly reduced size and cost of the receiving equipment at the ground station (which is usually the users premises or the handset) and superior modulation, coding and synchronization techniques. With the discovery of turbo codes, the aforementioned scenario has become a reality.In order to further improve the performance and bring down the cost of the receivers, we propose offset QPSK as the modulation technique, which allows the use of power efficient non-linear amplifiers and dataaided synchronization algorithms which have a faster acquisition time than the non-data-aided counterparts proposed in [1]. We also use the upsampled version of the matched filter as an interpolator [2], enabling the implementation of a feedforward timing acquisition method, that is faster than the feedback approach discussed and an all-digital implementation of carrier synchronization for digital radio systems is proposed in [18]. A comparison of different carrier recovery techniques is presented in [19]. A carrier recovery algorithm for M -ary QAM with a capability to track large frequency offsets is discussed in [20]. A non-data-aided carrier recovery method for modified 128-QAM is proposed in [21].Timing recovery can be broadly classified into synchronous and asynchronous methods. In the synchronous methods, the local clock at the receiver is regenerated from the incoming signal. Such techniques are implemented in hardware and are usually employed in analog modems [22]. In the asynchronous approach, the local clock at the receiver is free running, due to which the ratio of the receiver sampling frequency to the incoming symbol-rate is not an ...

Data Detection in Single User Massive MIMO Using Re-Transmissions

“…Joint ML detection and channel estimation in multiuser massive MIMO is presented in [38]. Detection of turbo coded offset QPSK in the presence of frequency and clock offsets and AWGN is presented in [39], [40]. Channel estimation in large antenna systems is given in [41]- [43].…”

Enhanced data‐aided frequency estimation by collaboration in a distributed receiver

“…In the simulated results, a strong receive node achieves BER of $$2\times 10^{-4}$$ at $$\rho = 8$$ dB when no frequency error is present and the distributed receiver (with one strong and 9 weak nodes) achieves the same BER at $$\rho = 2$$ dB, which clearly depicts 6 dB SNR gain. In addition to this, we can show the $$x$$‐axis in terms of average SNR per bit as shown in [25, 26].…”

Near Capacity Signaling over Fading Channels using Coherent Turbo Coded OFDM and Massive MIMO

“…where ĥk, m, nr, nt is obtained from (27). The fine frequency offset estimate (ν k, f (n r , n t )) is obtained by dividing the interval [ω k − 0.005, ωk + 0.005] radian (ω k is given in ( 20)) into B 2 = 64 frequency bins [44]. The reason for choosing 0.005 radian can be traced to Figure 5 of [3].…”

“…4) Cancel the frequency offset by multiplying rk, n, nr in ( 15) by e −j(ω k +ω k, f )n , and estimate the channel again using (27), for each n r and n t . 5) Obtain the average superfine frequency offset estimate using (44). Cancel the offset by multiplying rk, n, nr in ( 15) by e −j(ω k +ω k, f +ω k, sf )n .…”