J. Nossek

Technische Universität München, Department of Electrical Engineering and Information Technology, Faculty Member

Followers

Following

Co-authors

Public Views

Interests

Uploads

Papers by J. Nossek

Rate Balancing in Multiuser MIMO OFDM Systems

IEEE Transactions on Communications, 2009

Recently, the capacity region of the Gaussian broadcast channel has been characterized. For a giv... more Recently, the capacity region of the Gaussian broadcast channel has been characterized. For a given transmit power constraint, those points on the boundary of the capacity region can be regarded as the set of optimal operational points. The present work addresses the problem of selecting the point within this set that satisfies given constraints on the ratios between rates achieved by the different users in the network. This problem is usually known as rate balancing. To this end, the optimum iterative approach for general MIMO channels is revisited and adapted to an OFDM transmission scheme. Specifically, an algorithm is proposed that exploits the structure of the OFDM channel and whose convergence speed is essentially insensitive to the number of subcarriers. This is in contrast to a straightforward extension of the general MIMO algorithm to an OFDM scheme. Still, relatively high complexity and the need of a time-sharing policy to reach certain rates are at least two obstacles for a practical implementation of the optimum solution. Based on a novel decomposition technique for broadcast channels a suboptimum non-iterative algorithm is introduced that does not require time-sharing and very closely approaches the optimum solution.

Download

ZF Detectors over Correlated K Fading MIMO Channels

IEEE Transactions on Communications, 2011

This paper provides a systematic characterization of Zero-Forcing (ZF) detectors over multiple-in... more This paper provides a systematic characterization of Zero-Forcing (ZF) detectors over multiple-input multiple-output (MIMO) channels that experience both small and large-scale fading. In particular, we consider the generic distribution (Rayleigh/gamma distribution) to model the composite fading fluctuations and also assume the general case of semi-correlated small-scale fading. In the following, novel exact analytical expressions for the achievable sum rate are derived, followed by asymptotic expressions in the high and low Signal-to-Noise ratio (SNR) regimes. In these limiting cases, two common and insightful affine expansions are studied followed by new, closedform upper and lower bounds on the sum rate that remain tight for all SNRs. In the second part of the paper, we present exact tractable expressions along with first-order expansions for the symbol error rate (SER) and outage probability; we also quantify the performance of ZF detectors in terms of diversity order and array (or coding) gain. The implications of the model parameters on the ZF detector performance are investigated via Monte-Carlo simulations which also validate the theoretical analysis.

Download

Region 8: Importance of basic principles

IEEE Circuits and Systems Magazine, 2009

Modulation techniques for microwave digital radio

IEEE Communications Magazine, 1986

Download

New test results of a 4 by 4 discrete-time cellular neural network chip

CNNA '92 Proceedings Second International Workshop on Cellular Neural Networks and Their Applications

An analog implementation of discrete-time cellular neural networks based on conductance multiplie... more An analog implementation of discrete-time cellular neural networks based on conductance multipliers and transconductance amplifiers is discussed. The circuit architecture takes advantage of differential currents suppressing DC disturbances. Measurement results that compare the fabrication tolerances between different cells on the chip and the offset of the single cells are given

Representatives IEEE Solid-State Circuits Society IEEE Neural Networks Society IEEE Sensors Council

Q-enhancement and extension of the stability range of generalized immittance converters

IEEE Transactions on Circuits and Systems

A compensation technique is proposed which will extend the useful bandwidth of Antoniou-type immi... more A compensation technique is proposed which will extend the useful bandwidth of Antoniou-type immittance converters for the realization ofRC-active filters with high pole-Qtowards higher frequencies. Using operational amplifiers with very high unity-gain-frequency, stability problems are likely to occur. A method for an essential improvement of the stability margin is presented.

Analysis of a near field MIMO wireless channel using 5.6 GHz dipole antennas

2016 ESA Workshop on Aerospace EMC (Aerospace EMC), 2016

Understanding the impact of interference upon the performance of a multiple input multiple output... more Understanding the impact of interference upon the performance of a multiple input multiple output (MIMO) based device is of paramount importance in ensuring a design is both resilient and robust. In this work the effect of element-element interference in the creation of multiple channels of a wireless link approaching the near-field regime is studied. The elements of the 2-antenna transmit- and receive-arrays are chosen to be identical folded dipole antennas operating at 5.6 GHz. We find that two equally strong channels can be created even if the antennas interact at sub-wavelength distances, thus confirming previous theoretical predictions.

Built-in self test architectures for multistage interconnection networks

Proceedings ED&TC European Design and Test Conference, 1996

A novel built-in self test architecture for locally controlled cube-type N N multistage interconn... more A novel built-in self test architecture for locally controlled cube-type N N multistage interconnection networks (MINs) is presented. First, a state-based pseudoexhaustive test procedure for this class of MINs is outlined. Then, a labelling algorithm on a binary n-cube is described which generates the necessary inputs for the tests. From the dependence graph of this algorithm a tree architecture is derived which results in a hardware overhead of O1= log N.

Download

Multiwavelet transforms based on several scaling functions

Proceedings of IEEE-SP International Symposium on Time- Frequency and Time-Scale Analysis, 1994

An algebraic method for the design of discrete wavelet transforms based on several scaling juncti... more An algebraic method for the design of discrete wavelet transforms based on several scaling junctions is presented. Solving systems of partly nonlinear equations is necessary to compute the discrete Coefficients. Wavelet transforms based on several scaling junctions enable properties that are impossible an the single-wavelet case. Wavelet transforms based on several scaling junctions can also be designed wavelet-like, what leads to a better approximation of the continuous bases. In this paper we show how to construct the discrete wavelet transforms based on several scaling junctions with the algebraic design method and discuss the properties of the resulting wavelet bases.

Download

Simulation of electromagnetic effects in SAW RF filters

1997 IEEE MTT-S International Microwave Symposium Digest, 1997

The characteristics of surface acoustic wave (SAW) filters are influenced by the electromagnetic ... more The characteristics of surface acoustic wave (SAW) filters are influenced by the electromagnetic properties of the immediate environment of the acoustically active structures. This paper presents a model for electromagnetic effects caused by the filter environment. Here, the filter environment comprises the setup surrounding the SAW structures, e.g., pads, bonds, pins, etc. The components, which are particularly relevant to the performance of SAW filters, are identified. Measurement and simulation are compared showing good agreement with the characteristics of the filter

Retiming of synchronous circuits with variable topology

Proceedings of the 8th International Conference on VLSI Design, 1995

Generally, circuit design leads t o a trade-off scenario between speed and various parameters lak... more Generally, circuit design leads t o a trade-off scenario between speed and various parameters lake power dissapalion, AT complexity, re-use of already existing cells, design time, etc. To deal with this trade-off, the interaction between retiming and the selection of combinational elements from a set of cells with these different parameters is considered here. Additionally, modifications of the circuat graph concerning the interconnections, e.g. due to associativity of the underlying algorithm, lead to a parameterized topology. The algorithm presented in this paper combines all three, retiming, the selection of specijiic cells and the choice of an appropriate topology an one optimization step.

format_quoteThe paper pioneers a methodology merging retiming with circuit modification for global performance optimization, addressing complex design trade-offs.format_quote

Download

Continuation-based learning algorithm for discrete-time cellular neural networks

Proceedings of the Third IEEE International Workshop on Cellular Neural Networks and their Applications (CNNA-94), 1994

The SGN-type nonlinearity of a standad Dismte-The Cellular Neuml Network (DTCNN) is replaced by a... more The SGN-type nonlinearity of a standad Dismte-The Cellular Neuml Network (DTCNN) is replaced by a smooth, SienOidcJ nonlinearity with variabk gain. T h m f m , the malting dynamical system is fully diffmtiobk. Bounds on gain of the function are given, ao that the new, anaoth system appnmimatea the a t a d a d DTCNN d i n certain limits. A learning aighthm is prqposed, which findr the template pammeters for the standard DTCNN by & d I y increasing the gain of the sigmoidal finction.

Download

Low power CORDIC implementation using redundant number representation

Proceedings IEEE International Conference on Application-Specific Systems, Architectures and Processors, 1997

In this paper a niethodolog! for reclucing the power consumption of shift-and-add operations in g... more In this paper a niethodolog! for reclucing the power consumption of shift-and-add operations in general and especially of CORDIC stages is presented. The proposed method uses the fuct of simultaneous c a r p generation in rediiiidarit carp-save and signed digit structures to predict the miriimiini necessap hardware effortfor shift-and-add operations. As a carry once generated in a certain hit position carinot "ripple I' through the adder if using redundant number representation, hurdwnre purty can De switched oti or off depending on the shi8 constant. Simulations have shown, that shift dependent hardkvare iitilization ofparallel implementations leads to monotonicallj decreasing power consumption for increasing shift constants. A CORDIC processor element fiir 16 digit SDNR has heeri iniplemented as a layout and simulated with PowerMill in terms of power consumption.

Download

Multistage MMSE decision feedback equalization for EDGE

International Conference on Acoustics, Speech, and Signal Processing, 2003

Compared to wired channels, time dispersive radio channels possess more frequent s in their spect... more Compared to wired channels, time dispersive radio channels possess more frequent s in their spectral characteristics. Thus, the performance of linear filters to compensate intersymbol interference degrades dramatically. The well-known nonlinear decision feedback equalizer (DFE) is one approach to improve this behavior. However, in systems with observations of high dimensionality, the optimum DFE structure is computational intensive. In this paper,

Download

Outdated Uplink Adaptation Due to Changes in the Scheduling Decisions in Interfering Cells

2008 IEEE International Conference on Communications, 2008

Mario Castaneda, Michel T. Ivrlac and Josef A. Nossek Lehrstuhl für Netzwerktheorie und Signalve... more

Cross layer optimization - an equivalence class approach

ITG Workshop on Smart Antennas (IEEE Cat. No.04EX802), 2004

Performance gains of MIMO signal processing

e & i Elektrotechnik und Informationstechnik, 2005

In order to wisely use the degrees of freedom offered by a multi-input multi-output (MIMO) system... more In order to wisely use the degrees of freedom offered by a multi-input multi-output (MIMO) system it is necessary to identify and quantify the possible gains in performance one can expect from MIMO signal processing. In this paper, we derive measures which allow to quantify the amount of antenna gain, diversity gain, and multiplexing gain of a MIMO system. It turns out that it is impossible to maximize all three gains, or any pair of two gains at the same time. The necessary trade-off between performance gains is demonstrated.

Download

Low power register file architecture for application specific DSPs

2002 IEEE International Symposium on Circuits and Systems. Proceedings (Cat. No.02CH37353), 2002

In this paper, an architecture for register files suited for synthesiz- able DSP cores is propose... more In this paper, an architecture for register files suited for synthesiz- able DSP cores is proposed. The principal focus is on the imple- mentation of DSP algorithms with several identical channels, used in e.g. stereo audio, filterbanks or network IC implementations. Nevertheless, it is shown that the result of this work can be ex- tended to many single channel applications

Lossless cellular neural networks

1996 Fourth IEEE International Workshop on Cellular Neural Networks and their Applications Proceedings (CNNA-96), 1996

S i n c e their introduction, Cellular Neural Networks [4] have turned out to be useful architect... more S i n c e their introduction, Cellular Neural Networks [4] have turned out to be useful architectures for the solution of many problems, e. g. in image processing or in the simulation of partial diflerential equations. Therefore, there have been several attempts to introduce cell circuits suitable for large-scale integration [. ?I. Up to now, all of these cells need energy and therefore power supply. Just recently attempts have been made to build up circuitry being able to work without an external energy supply by using the energy stored in the initial state [I]. This principle can provide two major advantages. First, since no or at least not much energy is dissipated during computation, the circuit does not produce much heat. Therefore, there are no "hot spots" in integrated circuits, which limit integration density and operation speed. Furthermore, since there is no need for a power supply, the absence of voltage supply lines supports a high integration density. In this work an architecture for the realisation of a lossless C N N is proposed. Further on, since standard learning algorithms turn out to fail for lossless systems, a way to amend these is introduced.

Download

J. Nossek

Uploads

Papers by J. Nossek

Log In