Heinz G. Göckler

A comprehensive survey of digital transmultiplexing methods

With this survey, an attempt is made to describe the great majority of all known methods of digital transmultiplexing (i.e., conversion) of time-division-multiplex (TDM) to frequency-division-multiplex (FDM) signals, and vice versa. To this end, the individual transmultiplexer approaches are classified into four categories according to the underlying algorithm: Bandpass filter bank, low-pass filter bank, Weaver structure method, and multistage modulation method. Finally, the overall performance of the various transmultiplexer approaches are compared with each other by means of different criteria [1], such as stability under looped conditions, absolute value of the group delay, computational and control complexity, modularity, potential of intelligible crosstalk, absence of an additional analog frequency conversion, and the impact of out-of-band signaling For a more profound understanding of the individual digital trsnsmultiplexer approaches, the main chapter is preceded by an introductory discussion on analog and digital generation of single-sideband signals. In this context, the associated problems of sample rate alteration and multi-rate filtering arising from digital signal processing are dealt with.

Channel-individual adaptive beamforming for mobile satellite communications

A digital frequency-division multiplex demultiplexer for single channel per carrier (SCPC) signals was developed for use in a digital channel-individual beamforming network for an adaptive satellite array antenna for mobile satellite communications. The available ASIC (35000 gate functions in CMOS technology, clock rate up to 30 MHz with low power consumption) permits the separation of FDM signals comprising up to 16 channels. Starting from a suitable model for the demultiplexer beamforming system, this paper describes the analysis, design, simulation and implementation of the hierarchically structured demultiplexer. Aspects of digital beamforming are only discussed as far as necessary for understanding. >

Dynamic Frequency-Band Reallocation and Allocation: from Satellite-Based Communication Systems to Cognitive Radios

This paper discusses two approaches for the baseband processing part of cognitive radios. These approaches can be used depending on the availability of (i) a composite signal comprising several user signals or, (ii) the individual user signals. The aim is to introduce solutions which can support different bandwidths and center frequencies for a large set of users and at the cost of simple modifications on the same hardware platform. Such structures have previously been used for satellite-based communication systems and the paper aims to outline their possible applications in the context of cognitive radios. For this purpose, dynamic frequencyband allocation (DFBA) and reallocation (DFBR) structures based on multirate building blocks are introduced and their reconfigurability issues with respect to the required reconfigurability measures in cognitive radios are discussed.

Minimal block processing approach to fractional sample rate conversion

Abstract The problem of synchronous fractional sample rate conversion (FSRC) of a digital signal by L / M , where L and M are coprime integers [6,10], is revisited. Based on a novel approach two different efficient causal block implementations of FSRC are concurrently derived and compared with each other. While the computational load of both structures, being performed in an LTI MIMO subsystem at the subnyquist rate F o / L = F i / M , is identical, their group delay is always different. By column and row shifts of the matrix representation of the MIMO subsystem it is possible to transform each structure into any arbitrary implementation with changed group delay. Moreover it is shown that, by structural manipulation of the signal flow graph of the MIMO subsystem, both implementations ultimately require the same amount of computation and storage in spite of different group delay. Finally, by using Nyquist( L )filters, the maximum number of input samples to FSRC is retained at its output.

A Modular Approach to a Digital 60-Channel Transmultiplexer Using Directional Filters

A multistep transmultiplexer approach with a single-way modulation scheme applying minimum phase wave digital directional or transversal filters is proposed. Furthermore, the impact of signaling, pilots, and spectral shaping at the PCM end is investigated. The results of a filter design based on wave digital filters is given. The main merits of this approach are, at a moderate level of computational complexity, its absolute stability, the low basic group delay, and its modularity. Due to modularity, the central unit (the TDM/FDM and FDM/TDM translators) is composed of only three different filter types requiring neither DFT, FFT, or associated processors nor modulators. As a consequence, the new approach is suitable for LSI/VLSI integration with the potential of efficient testability.

A novel approach to the design of oversampling low-delay complex-modulated filter bank pairs

In this contribution we present a method to design prototype filters of oversampling uniform complex-modulated FIR filter bank pairs. Especially, we present a noniterative two-step procedure: (i) design of analysis prototype filter with minimum group delay and approximately linear-phase frequency response in the passband and the transition band and (ii) Design of synthesis prototype filter such that the filter bank pairs distortion function approximates a linear-phase allpass function. Both aliasing and imaging are controlled by introducing sophisticated stopband constraints in both steps. Moreover, we investigate the delay properties of oversampling uniform complex-modulated FIR filter bank pairs in order to achieve the lowest possible filter bank delay. An illustrative design example demonstrates the potential of the design approach.

Tree-structured MIMO FIR Filter Banks for Flexible Frequency Reallocation

Emerging digital on-board processors for communication satellites concurrently call for two tasks of a cascade of a frequency demultiplexer (FDMUX) and a frequency multiplexer (FMUX): i) mere de-and remultiplexing of an FDM signal (FDFMUX functionality) and ii) decomposition and reconstruction of a wideband signal (Sub-band Coding (SBC) functionality). In this paper, two FIR approaches allowing for flexible frequency reallocation, the oversampling 2times2-channel and the critically sampling 2-channel FIR SBC-FDFMUX filter banks, are investigated and compared with each other. Furthermore, tree-structured MIMO extensions of both approaches including the application of switching functions are investigated. The oversampling 2times2-channel SBC-FDFMUX filter bank moderately outperforms the critically sampling 2-channel SBC-FDFMUX filter bank in terms of computational complexity, whereas the latter is highly modular and, hence, simply realisable.

DESIGN OF RECURSIVE POLYPHASE NETWORKS WITH OPTIMUM MAGNITUDE AND MINIMUM PHASE

A novel approach to the design of recursive polyphase networks with application to multirate filtering and digital filter banks is proposed. To this end, Pellonis method (4) is extended such that all zeros of the polyphase network are located on the z-plane unit circle and all poles are single, corresponding to an optimum magnitude response with minimum phase. An example, taken from transmultiplexing applications, shows that the proposed method results in substantially improved filter designs compared with the approaches of Bellanger et al. (l), Vary (6) and Pelloni (4). Zusarnrnenfassung. Es wird ein verbessertes Verfahren zum Entwurf von rekursiven Polyphasennetzwerken vorgeschlagen. Zu diesem Zweck wird die von Pelloni (4) angegebene Methode so rnodifiziert, dafl alle Nullstellen des resultierenden Polyphasenetzwerks auf der Peripherie des Einheitskreises in der z-Ebene liegen und alle Pole einfach sind, was zu einem optimalen Betragsverlauf und zu minimaler Phase der Filterwirkungsfunktion fiihrt. Das Beispiel eines Zweikanaltransmul- tiplexers zeigt, dafl man mit dem vorgeschlagenen Verfahren Entwiirfe erhalt, die im Vergleich zu den Ergebnissen von Bellanger et al. (l) and Vary (6) erheblich weniger Aufwand erfordern. Gegeniiber der Methode von Pelloni ergeben sich deutliche Verbesserungen des Frequenzgangs im Durchlaflbereich.

Parallelisation of Digital Signal Processing in Uniform and Reconfigurable Filter Banks for Satellite Communications

Modern satellite communication calls for novel and flexible concepts for de- and remultiplexing of wide-band FDM-signals in conjunction with beam switching on-board a satellite. For this application, two highly efficient approaches to uniform and yet reconfigurable digital filter banks are known: i) the complex-modulated polyphase filter bank, and ii) the tree-structured filter bank. Channelising and remultiplexing of (ultra-)wide-band FDM-signals require high end sample rates that, often, range beyond technological limit (e.g. > l GHz). A remedy is to process the signals at a lower rate in a parallel manner. While the polyphase approach i) applying one-step sample rate alteration (down/up-sampling) inherently operates in parallel, the hierarchical approach ii) being based on stepwise sample rate alteration calls for the parallelisation of, at least, the high rate stages. In this contribution, the systematic sample-by-sample processing procedure for parallelisation of multirate systems by Groth is recalled. It is applied to the parallelisation of a novel class of tree-structured filter banks: The various steps towards parallelisation of the high rate front end of the FDM-demultiplexer part of the filter bank are described in detail

Efficient minimum group delay block processing approach to fractional sample rate conversion

The derivation of block-processing structures for fractional sample rate conversion (FSRC) is revisited. In the past, several time- or frequency-domain approaches have been proposed. However, the resulting structures differ concerning their group delay and, as a consequence thereof, in the arrangement of coefficients. In this paper it is shown that a minimum extra delay of z/sup -(ML-L-M+1)/ is sufficient to ensure causality. This outcome is demonstrated both in frequency- and time-domain.