Michael Bergmann

Channel Estimation on the Forward Link of Multi-beam Satellite Systems

Multi-beam concepts are an essential component of next generation broadband satellite systems. Due to aggressive design goals, full frequency reuse is suggested in this context so that appropriate interference mitigation techniques have to be applied on both forward and return links. In this respect, channel estimation is of paramount importance. Throughout this paper, we are focusing on channel estimation of the forward link with emphasis on orthogonal and non-orthogonal training sequences used for this purpose. By analytical and simulation results, it is confirmed that the former are best suited in terms of the obtained jitter performance. On the other hand, non-orthogonal codes are not restricted by their length, but it is shown that a linearly independent set of unique words is significantly affected by an amplification of the noise component – a result not available from the technical literature on this subject so far. Furthermore, it is demonstrated that a simple correlation procedure, which might be employed for any kind of non-orthogonal training sequences, produces a non-negligible jitter floor in the higher SNR regime, which is primarily given by the cross-correlation properties of the code.

Code-aware joint estimation of carrier phase and SNR for linear modulation schemes

Depending on the length of the training sequence, data-aided parameter estimation requires some extra bandwidth. On the other hand, with non-data-aided solutions the payload symbols can be exploited as well, although the performance deviates significantly from the Cramer-Rao lower bound (CRLB) in the low signal-to-noise ratio (SNR) regime. In order to bridge this gap, recent activities concentrated more and more on code-aware (CA) algorithms so as to use the potential of error correction schemes, which usually form part of modern communication systems. In this context, methods for both carrier phase and SNR estimation have been published in the open literature. In the current paper, we extend this work by developing a CA algorithm for joint estimation of carrier phase and SNR; for comparison purposes, the CRLB of this joint approach is derived as the theoretical performance limit.

New Results on Location-Aware Channel Estimation for Multibeam Satellite Links

In order to employ interference mitigation techniques in a multibeam satellite scenario under full frequency reuse, channel estimation is of paramount importance. Even on symbol-synchronous forward links, this is a most challenging task since it must be accomplished before mitigation techniques are applicable. Pilot-aided sequences are normally used in this context, where problems with interfering signals are avoidable, if such sequences satisfy the orthogonality criterion. In case that nonorthogonal schemes have to be employed because of format and/or efficiency reasons, the estimator performance degrades in part tremendously by noise amplification or self-noise effects. Motivated by this background, we derive a maximum-likelihood solution by taking into account the location of the user terminals on ground. A significant improvement of the error performance is achieved with this approach by modifying the interfering components in terms of amplitude and phase such that they constructively contribute to the useful signal part.

Parameter estimation for link adaptation on land-mobile satellite links

Propagation conditions on a land-mobile satellite (LMS) link may change within the round-trip time. While for slowly fading channels the signal-to-noise ratio (SNR) is the most important figure of merit to adapt the related transmission parameters, this does not hold true for mobile channels in general. Additional information about the signal-to-interference ratio (SIR) and the Doppler spread of the fading component are typically necessary for successful link adaptation techniques. In this paper, we address carrier recovery as well as estimation of SNR, SIR, and Doppler spread for a Loo-Fontan channel, which is in the open literature frequently considered as a suitable model for LMS links. The focus is on the performance of an estimation framework, which we identified in a precursor study to be appropriate for correlated Rice channels. This developed framework is extended to LF conditions and verified by simulation results.

Estimation of carrier and channel parameters for land-mobile satellite channels

Information about the channel state is particularly useful for adapting transmission parameters to different fading conditions. With slowly fading channels, the signal-to-noise ratio (SNR) is the most significant figure of merit to obtain the required goals in terms of availability and throughput. Using mobile satellite links, however, the situation may change within the round-trip time. Therefore, not just SNR estimates have to be delivered to the transmitter station in the context of an adaptive coding and modulation strategy, but also some additional information about the signal-to-interference ratio (SIR) and, perhaps more important, the Doppler spread of the mobile link, which determines the amount of variation in time. Assuming a channel determined by correlated Rician fading, an appropriate estimation framework for carrier frequency and phase, SNR, SIR, and Doppler spread will be discussed in the current paper. The signal model reflects in principle the conditions on a time-selective land-mobile satellite link, whereas frequency selectivity - as it is typical for multipath propagation - is considered to be negligible.

Generic communication user and system requirements for future space science missions

Since space missions have been flown, a proper data return to Earth was inherently required. The complexity level and the number of space experiments have increased over the last decades. As a consequence, higher data rates have been required. The most limiting factor on achievable data rates for interplanetary communications is the distance. It imposes challenges on all parts of the communication system. Communication user and system requirements must be carefully defined to fulfill the mission needs. These requirements are expected to increase in the future. Hence, we studied 13 generic future ESA space mission scenarios comprising 5 Lunar missions, 5 Mars missions, and three missions to special targets (Lagrange point L2, NEO, and the Jovian satellite Europa) based on planned and already flown missions in order to derive communication user and system requirements. These requirements were classified, scaled, adapted and validated by calculations for these missions defined in [1]. Within this paper we will present how communication user and system requirements for future space science missions can be classified, scaled and applied. Therefore, after introducing a proper classification of communication requirements we will present recommendations by means of concrete examples assuming an Asteroid Sample Return mission scenario. Some practical hints will be provided as well.

Joint recovery of carrier frequency and symbol timing for extremely bandwidth-efficient satellite links

Efficient estimation and synchronisation of transmission parameters is key for any satellite receiver, which is especially true for the recovery of carrier frequency and symbol timing. In this respect, a joint feedback solution has been analysed and tested in some recent work conducted by the authors. However, it turned out that the capture range of the carrier frequency is limited by the selected roll-off factor (excess bandwidth), which means that the synchroniser would not work properly for extremely bandwidth-efficient transmission schemes. This motivated them to modify the feedback structure by introducing a different filter operated in parallel to the receiver matched filter. In the sequel, it could be verified by analytical and simulation results that the initial frequency offset synchronised successfully can reach up to ±50% of the symbol rate irrespective of the selected modulation scheme, the roll-off factor, the signal-to-noise ratio value and the timing offset the satellite link is operated with. On top of that, a non-linear modification of the timing error detector has been investigated in detail, since the output of the conventional solution is nullified in case of extremely small excess bandwidths.

Estimation of carrier and channel parameters in time-selective fading channels

Assuming a land-mobile satellite link with negligible multipath and shadowing effects, the corresponding channel model is adequately described by a correlated Rician distribution. In this context, the estimation of carrier and channel parameters has been investigated in some earlier work by means of an algorithmic approach primarily based on heuristic considerations. In the current study, the authors will review the results obtained so far from a maximum-likelihood (ML) point of view. A major output of the analysis is that the simple ad-hoc framework achieved previously is indeed an ML solution, when the impairments introduced by Rician fading are spectrally flat and strictly band limited.

Robustness analysis of location-aware channel estimation on multibeam satellite links

A total throughput of about 100 Gbit/s is state-of-the-art for current communication satellite systems. But next generations are envisaged to improve this by a factor of ten, as postulated by the European Space Agency (ESA) in a recent project, identifying multibeam architectures in combination with an aggressive reuse of frequency bands as crucial enabler technologies. Interference problems, related to this approach, could be mitigated by (joint) precoding and beamforming on the forward link and multi-user detection on the return link. However, these methods necessitate accurate and timely information upon the channel state. In this context, just by exploiting the a priori knowledge about the antenna characteristics and the user terminal position, a novel strategy called location-aware channel estimation (LACE) was published recently; promising results in terms of accuracy and throughput were achieved in this context. The current paper examines and demonstrates the robustness of the LACE concept in case the user position is not perfectly known.