No abstract

In this paper, we review the key figures of merit to assess the performance of advanced random access (RA) schemes exploiting physical layer coding, repetitions and collision resolution techniques. We then investigate RA modeling aspects and their impact on the figures of merit for the exemplary advanced RA schemes: Contention Resolution Diversity Slotted ALOHA (CRDSA), Irregular Repetition Slotted ALOHA (IRSA), Coded Slotted ALOHA (CSA) and Enhanced Spread-Spectrum ALOHA (E-SSA). We show that typical simplifications of the reception model when used to optimize RA schemes lead to inaccurate findings, both in terms of parameter optimization and figures of merit, such as the packet loss ratio (PLR) and throughput. We also derive a generic RA energy efficiency model able to compare the schemes in terms of the energy required to transmit a packet. The combination of achievable RA throughput at the target PLR and energy efficiency, for the same average user power investment per frame and occupied bandwidth, shows that E-SSA, which is an unslotted scheme, provides the best overall performance, while, in terms of the slotted schemes, CRDSA outperforms the more elaborated IRSA and CSA. This surprising results is due to the fact that the IRSA and CSA optimization has so far been performed using RA channel models that are not accurately reflecting the physical layer receiver behavior. We conclude by providing insights on how to include more accurate reception models in the IRSA and CSA design and optimization.

Precoding for multibeam satellites with aggressive frequency reuse has attracted interest of late towards enhancing system capacity. Most of the works on precoding mitigate the linear co-channel interference between the beams caused by frequency reuse. However, the high power amplifier (HPA), an integral part of the satellite payload, is inherently non linear. Non-linear amplification combined with the linear co-channel interference introduces nonlinear co-channel distortions at the receiver. Further, signals with very high peak to average power ratios (PAPR), typical of spectrally efficient modulations, are sensitive to the non-linear characteristic of the HPA and necessitate large back-off to have manageable distortion levels. In this work, a novel architecture comprising multistream Crest Factor Reduction (signal pre-clipping) and Signal Predistortion (SPD) in cascade, is devised to counter the nonlinearities and co-channel interference in multibeam satellite systems. An iterative algorithm to optimize the parameters of the signal clipping and predistortion is devised taking recourse to analytical derivations. The proposed joint estimation paradigm is shown to compare favorably with state-of-art and provides a framework to combine predistortion and precoding.

This paper studies the End-to-End (E2E) performance of a High Throughput Satellite (HTS) system that utilizes a transparent (non-regenerative) optical feeder link to transport the DVB-S2 signals associated with the large number of spotbeams in the radio access link. Two main sources of nonlinearity are considered to characterize the effect that the satellite forward link has in the DVB-S2 signal, namely the optical Mach-Zehnder Modulator (MZM) in the gateway and the High Power Amplifier (HPA) in the satellite. Digital Pre-Distortion (DPD) and linear equalization are also implemented to mitigate the impact of the non-linear distortion and the inter-symbol interference on the Packet Error Rate (PER) of the received DVB-S2 signal. The performance analysis is carried out using the Total Degradation (TD) as performance indicator. As expected, the Input Back-Off (IBO) of the HPA and the modulation index of the MZM should be jointly optimized according to the DVB-S2 MODCOD that is selected for transmission. Therefore, the stronger is the Forward Error Correction (FEC) code that is used, the lower is the IBO and the higher is the modulation index that can be utilized to minimize the impact of non-linearities in the E2E performance.

No abstract