End-to-End Performance Evaluation of SLP Waveforms

In this paper, we investigate the demodulation and decoding of the symbol-level precoded (SLP) signal using the conventional logarithmic likelihood ratio (LLR) and low-density parity check (LDPC) decoding algorithms operating in real-time on the software-defined radio (SDR) receiver. Its conventional forward error correction (FEC) is designed assuming symmetric additive white Gaussian noise (AWGN). While SLP generates asymmetric constellation point clouds it is unclear if it can achieve similar FEC gains in an end-to-end coded setup. The receiver is implemented based on the DVB-S2X standard. It estimates multi-beam interference-limited channel coefficients by using the orthogonal pilots embedded into the DVB-S2X frames to facilitate precoding at the transmitter. After the received frames are fully synchronized the payload symbols are decoded in two stages: soft decision using the LLR and hard decision using the LDPC decoders. We benchmark the bit error rate (BER) and frame error rate (FER) of the complete decoding chain while using the zero-forcing (ZF) and the selected SLP channel-interference canceling techniques. The conducted in-lab experiments show that the SLP can improve BER and FER performance in the conventional receiving and decoding design for symmetric constellations while achieving its original optimization goals.