Arne Schulz

System level optimization and design space exploration for low power

We present a software tool for power dissipation analysis and optimization on the algorithmic abstraction level from C/C++ and VHDL descriptions. An analysis is most efficient on such a high level since the influence of design decisions on the power demand increases with increasing abstraction (A. Raghunathan et al., 1998). The ORINOCO/sup (R)/ tool enables us to compare different but functionally equivalent algorithms and bindings to RT-level architectures with respect to power consumption. The results of the optimized binding can be used to guide synthesis. In the experimental evaluation we compare the predicted optimization trend with synthesized implementations and prove the accuracy of our methodology and tool.

A Service Oriented Platform for Health Services and Ambient Assisted Living

Providing health care and assistance at home becomes more and more important due to the aging society and - in general - a health system under financial pressure. It is generally accepted that these services have to be supported by an eHealth infrastructure that enables the exchange of patient related data between different health institutions and assistance systems athome (referred to by the term AAL). In this paper a service-oriented architecture for delivering eHealth / AAL services at home using a hardware platform such as a residential gateway or a set-top-box is proposedand exemplified by three different services: (1) telerehabilitation of patients after heard surgery, (2) support of hearing impaired people and (3) the monitoring of Activities of Daily Living (ADL).

An Improved Power Macro-Model for Arithmetic Datapath Components

We propose an improved power macro-model for arithmetic datapath components, which is based on spatio-temporal correlations of two consecutive input vectors and the output vector. Based on the enhanced Hamming-distance model [3], we introduce an additional spatial distance for the input vector and the Hamming-distance of the output vector to improve model accuracy significantly. Experimental results show that the models standard deviation is reduced by 3% for small components and up to 23% for complex components. Because of its fast and accurate power prediction, this model can be used for fast high-level power analysis.

Interconnect driven low power high-level synthesis

This work is a contribution to high level synthesis for low power systems. While device feature size decreases, interconnect power becomes a dominating factor. Thus it is important that accurate physical information is used during high-level synthesis [1]. We propose a new power optimisation algorithm for RT-level netlists. The optimisation performs simultaneously slicing-tree structure-based floorplanning and functional unit binding and allocation. Since floorplanning, binding and allocation can use the information generated by the other step, the algorithm can greatly optimise the interconnect power. Compared to interconnect unaware power optimised circuits, it shows that interconnect power can be reduced by an average of 42.7 %, while reducing overall power by 21.7 % on an average. The functional unit power remains nearly unchanged. These optimisations are not achieved at the expense of area.

A fitting method for headphones to compensate individual hearing impairments

Sensorineural hearing loss often results in altered loudness perception and a smaller perceivable dynamic range. To compensate sensorineural hearing loss it is necessary to fit an individual compressive gain. Fitting of hearing aids is usually done by professionals together with the hearing aid user. To give the users of TV-headsets the possibility to adjust not only the volume of the acoustic signal but also the dynamics, three different versions of interactive self-fitting systems were designed and evaluated. The user tests were performed with 11 subjects with mild to moderate hearing loss. The results were evaluated with regard to benefit in quailty, loudness perception and speech intelligibility.