Thomas Kurbiel

Learn More
This paper presents a novel digital MIMO filter bank technique for modern broadband multibeam communcations satellite systems, which efficiently exploit available bandwidth by providing different users with different bandwidths for different types of services. This necessitates a flexible and dynamic resource reallocation, where the assumed traffic is(More)
This paper addresses the minimisation of the overall disturbances of the output signal of an oversampling DFT subband-coder filter bank. These disturbances are separated into aliasing and imaging. By means of the definitions of these disturbance components, two quadratic functions are derived that are used as objective functions in the design approach. The(More)
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)(More)
Hearing instruments employ filter bank systems (FBS) to subdivide the microphone signal into subband signals. To reduce overall computation, decimation is applied, which inevitably induces disturbance. First, these effects are analysed in detail. Then, we postulate application specific FBS requirements in view of extensive subband signal amplification.(More)
Oversampling, complex-modulated digital subband coder filter banks are commonly adopted in modern hearing aids to allow for individual amplification of subband signals in order to compensate for hearing losses of impaired persons. Minimum power consumption and low group delay have been the main concern in the design of these filter banks. Guaranteeing(More)
In this contribution, the design of approximately linear-phase, low delay and quasi-equiripple FIR lowpass filters is resumed with particular emphasis on very low group delay. To this end, a novel unconstrained single-objective multi-criterion design algorithm is proposed. Results show that the mean group delay can be diminished down to roughly 20%of that(More)
The vanishing gradient problem was a major obstacle for the success of deep learning. In recent years it was gradually alleviated through multiple different techniques. However the problem was not really overcome in a fundamental way, since it is inherent to neural networks with activation functions based on dot products. In a series of papers, we are going(More)
This paper addresses the improvement of the signal-to-disturbance ratio of the output-signal of an oversampling, complex-modulated subband-coder filter-bank pair with extensive subband-signal amplification. The undesired non-linear disturbance is induced by the sample rate conversion within the filter-bank pair and it is hard to predict in dependence of the(More)
  • 1