SERIALLY CONCATENATED ARTM TIER I WAVEFORMS WITH ITERATIVE DETECTION

We investigate the performance of Feher-patented quadrature phase-shift keying (FQPSK) and shapedoffset QPSK (SOQPSK) when serially concatenated with an outer code. We show that the receiver complexity for FQPSK and SOQPSK can be greatly reduced by viewing them as continuous phase modulation (CPM) waveforms. We use the pulse amplitude modulation (PAM) representation of CPM, which allows near-optimum detection of both modulations using a simple 4-state trellis. We compare the performance of the PAM-based approximation with another common approximation known as frequency/phase pulse truncation (PT). We use both of these reduced-complexity designs in serially concatenated coding schemes with iterative detection. In the end, we show that the PAM approximation has a slight performance advantage over PT, but both approximations achieve large coding gains in the proposed serially concatenated systems. INTRODUCTION In Offset QPSK (OQPSK), the quadrature component of the modulated carrier is delayed half a symbol time relative to the inphase component to avoid instantaneous 180◦ phase shifts. This is done to reduce the amount of spectral regrowth when a non-linear power amplifier is used. The spectral containment can be greatly improved by using cross correlated bandwidth efficient pulses, as is the case with Feherpatented QPSK (FQPSK) [2]. Shaped offset QPSK (SOQPSK) is very similar to FQPSK, except that the signal is typically viewed as a continuous phase modulation (CPM) rather than as a cross correlated linear modulation. Versions of these two waveforms, SOQPSK and FQPSK, have been incorporated into IRIG-106 as the “ARTM Tier I” waveforms [3]. When it comes to detecting the ARTM Tier I waveforms, the usual approach is to use an OQPSKtype detector. This simple (and suboptimal) detector is nothing more than a detection filter followed by a decision device. The primary drawback of this approach is that it ignores the inherent memory of these modulations. In this paper, we address this problem by using a strict CPM point of view when designing the detector. This is a very straightforward and natural approach for SOQPSK. On the other hand, FQPSK cannot be exactly represented as a CPM; therefore, a very close CPM approximation of FQPSK is used in place of the exact FQPSK model. The primary advantage of this CPM-based approach is that the memory of the signal is properly modeled in the detector. This leads to optimal or near-optimal detectors. Another advantage of the CPM-based approach is that the modulation itself can be viewed as a code. Therefore, the CPM-based detector is especially useful when codes are concatenated with the modulation and when iterative (turbo) detection is used. In the next section we describe the signal model used in the transmitters and receivers. In the section after that we show how the CPM model is used to construct simple, near-optimal detectors for SOQPSK and FQPSK. As we shall see, the CPM model alone is not sufficient to yield low-complexity detectors. Therefore, two well-known complexity-reduction techniques for partial-response CPM are used. The first is the pulse amplitude modulation (PAM) approximation [4], which was recently extended in [5] to ternary CPMs such as SOQPSK. The PAM technique allows the use of the simple 4-state trellis in [1] with a loss of only 0.1 dB in the uncoded case. The second reduced-complexity option is frequency/phase pulse truncation (PT) [6], which also allows the use of the simple 4-state trellis with a slightly larger loss of 0.2 dB in the uncoded case. The final objective for these reduced-complexity designs is in constructing simple decoders for serially concatenated coded FQPSK and SOQPSK. We show that the proposed CPM-based designs achieve large coding gains which are similar to those reported in [1] for serially concatenated coded military standard (MIL-STD) SOQPSK systems.