Standardization on performance objectives of satellite systems using adaptive coding and modulation

Abstract

Summary The radiocommunication part of the International Telecommunication Union (ITU‐R) establishes the standards for wireless communications. Recommendation ITU‐R S.1062 specifies the performance objectives for satellite communication systems providing constant bit rate service. However, this is not applicable to most modern satellite communication systems utilizing adaptive coding and modulation (ACM) schemes to compensate for channel impairments, especially for systems operating in high frequency bands. For this reason, ITU‐R recently developed Recommendation ITU‐R S.2131 that is suitable for satellite systems with ACM. The developed Recommendation provides a method to determine performance objectives for satellite systems using ACM. Two possible objective parameters are presented, including packet error rate and spectral efficiency. This paper presents the standardization process, and the background used to establish new performance objectives. In addition, interpretations of the objectives are given by using estimation results for two different propagation characteristics.

Adaptive modem and interference suppression based on deep learning

“…Adaptive code modulation (ACM) 2 and interference suppression play an important role in the wireless communication systems. Specifically, ACM can achieve a good compromise between capacity and reliability by adapting parameters to the channel condition 3 and has been applied in the cellular communication systems, 4 satellite broadcasting systems, [5][6][7] and optical communication networks. 8,9 However, traditional ACM usually needs amount of hardware and software resources to implement multiple codes and modems with different rates and different orders separately, and then select a certain combination of them to complete the communication according to the perception and decision of the channel environment.…”

“…But they mainly focus on a single modulation mode rather than a multi-modem by utilizing a unified deep learning architecture, a unified loss function, and a unified optimization algorithm. For example, In References 18,19, modems of multiple (n, k) combinations, including (2, 2), (4, 4), (7,4), (8,8), were realized by the DAN. However, the structure of the DAN had to be changed to adapt different combinations of (n, k).…”