Michel Bousquet

A programmable demultiplexer using a polyphase approach for regenerative satellites

This paper proposes a programmable demultiplexer architecture for the demultiplexing of carriers onboard a regenerative satellite. The demultiplxed carriers can have different bandwidths. The proposed architecture is a two stage one. The first stage uses a per channel approach and the second stage uses a polyphase approach. It has been demonstrated, through an example, that it is possible to program the demultiplexer to adapt it to different traffic configurations. Programming is performed at the software level. Development of satellite communication systems plans for onboard regeneration. This means that the uplink carriers are demodulated onboard of satellite. In comparison with conventionnal satellites, onboard regeneration offers several advantages : a gain of 2 to 5 dB on the link budget [MAR-861. independance of uplink and downlink. Different transmission modes can be adopted : for instance, frequency division multiple access (FDMA) for the uplink and time division multiplexing (TDM) for the downlink. The use of FDMA for the uplink allows low emission power and small antenna diameter for the earth stations. Dowlink TDM permits a higher utilisation of the satellite emission power because of zero intermodulation noise [CAM-88Al. with conventional satellites, VSAT networks are star-shaped with a hub station providing interconnectivity between all VSATs, as a result of limited link performance ; due to the increased link budget, onboard regeneration allows direct linkage between VSATs and hence avoids double hop. Transmission delay is divided by a factor of two. In spite of the above advantages, satellite operators hesitate somewhat to order regenerative satellites. This is essentially because of the operational rigidity which rises from the necessity to feed the satellite with carriers whose modulation format and arrangment in the transponder bandwidth are compatible with the processing that is done on board. This rigidity is a serious drawback in an operational and competitive context when compared to the flexibility offered by transparent payloads which accept any type of modulation and traffic. However this rigidity can be reduced to some extent by employing programmable onboard equipment. Demultiplexers are a key element for the onboard processing. Demultiplexing of a multicarrier signal can be performed in different ways : per channel approach ; multistage / tree approach ; ITT/IFFT approach ; polyphase approach. The per channel approach can be programmed by changing the coefficients of the filters and the decimation factor [DEL-881. Programming is flexible but the per channel approach is not suitable for the case where the number of carriers is important because the complexity of the demutiplexer is proportional to the number of carriers. The multistage / Tree approach can be made programmable, but it is only well adapted for the case where the number of carriers is a power of two. It is not considered in this paper. The FFT / IFFT approach is programmable [CAM88Bl. But this approach is computationnally efficient only for the case where the number of carriers is important (for example 1000 channels) [CAM-901. The polyphase approach is programmable, provided that one accepts that all carriers change bandwitdh equally Indeed the polyphase approach works for carriers of equal bandwidths [GAR-851 [DEL-881 [YIM-881. Nevertheless the polyphase approach is the most suitable candidate in terms of computation complexity where demultiplexing involves hundreds of carriers [CAM-901. Copyright @ 1991 by GUO, MARAL a d BOUSQUET. Published by the AIAA, Inc. with permission 1227 This paper proposes a polyphase based programmable demultiplexer, which overcomes the constraint of maintaining equal carrier bandwidth, thus allowing flexible reconfiguration of carriers. The total bandwith remains fixed, as is usually the case with a satellite transponder. The proposed programmable demultiplexer is a two stage one. The first stage uses a per channel approach and the second stage uses a polyphase approach. Hence it combines the high flexibility of the per channel approach and the low computation demand of the polyphase approach. The proposed architecture remains the same for different traffic configurations. The only changes needed are at software level. In section 2, a traffic model is presented. Section 3 presents the architecture of the demultiplexer. Section 4 discusses the implementation. Conclusions are summarized in section 5. The uplink signal consists of several carriers combined into a so-called multicarrier signal. The carriers are arranged in groups, with carriers of equal bandwidth within each group, but different bandwidths from one group to the other. This is not a severe operational constraint as in most practical situations, carriers are of standard formats, and a transponder very often displays a similar arrangement as that of of figure 1. Figure 1 illustrates a typical example with three types of carriers which correspond to the ISDN standard interface rates : 144 kbit/s (2B+D), 1544 kbit/s (23B+D) and 2048 kbit/s (30B+D). Traffics changes are assumed to occur as follows : the total bandwidth remains constant, the number of carriers within each group can vary, new carriers display no other format than the previously existing ones. The arrangement of figure 1 can be demultiplexed by the two stage demultiplexer architecture presented in figure 2. The first stage is composed of three filters which are used to separate the three groups of carriers. The second stage is a polyphase demultiplexer which is used to separate carriers of the same group. The three filters of the first stage are of the per channel type. Their transfer function is determined by the filter coefficients [CRO-831 [GAR-851 [DEL-901. They can be easily adapted to the required traffic pattern by selecting the proper values of the coefficients of filters and the decimation factor. The problem is how to adapt the polyphase demultiplexer of the second stage to the traffic

Reduced complexity frequency estimator for burst transmission

In burst transmission, carrier recovery is a critical point for synchronization systems. With a feedforward carrier phase recovery algorithm, a small frequency offset can significantly increase the cycle slip rate and then the phase error variance. Therefore, in order to obtain an accurate carrier phase estimation, a precise frequency correction is required. For M-states phase shift keying (M-PSK) modulated signals an unbiased feedforward reduced complexity frequency estimator (RCFE), operating in the non-data aided mode (NDA) is derived from the maximum likelihood (ML) principle. A compromise is realized between noise filtering and estimation slip probability by minimizing the estimator variance. It is optimized to operate at a low signal-to-noise ratio and short bursts. Its performance is compared to that of the ML estimator. The estimator is applied to an all-feedforward synchronization structure with QPSK modulated signals. Global performance of the modem synchronization structure is supplied.

Quelques applications des boucles à verrouillage de phase aux problèmes de synchronisation dans les liaisons spatiales

AnalyseCet article présente un ensemble de dispositifs assurant la synchronisation de porteuse ou de symbole dans des systèmes de télécommunications numériques spatiales, c’est-à-dire dans un domaine où le rapport signal à bruit est faible. Ces dispositifs fonctionnent en boucle fermée et réalisent de façon plus ou moins approchée une estimation de phase conforme au critère du maximum de vraisemblance. Sont successivement présentés et comparés entre eux les synchronisateurs de porteuse du type boucle à retour de décision, boucle de Costas et synchronisateur à non-linéarité en x2, et les synchronisateurs de symbole « early late gate », boucle de poursuite par transition de données et synchronisateurs à non-linéarité.AbstractThis paper deals with a presentation of carrier and bit synchronizers used for low signal-to-noise ratio space digital communications. Such synchronizers are closed loop phase estimators and carry out a mechanization of the maximum likelihood criterium. The paper presents well known carrier synchronizers such as the decision feedback synchronizer, the Costas loop and the x2 non-linearity synchronizer. Bit synchronizers such as the early late gate, the data transition tracking loop and others synchronizers associated with a non-linearity are also introduced.