Jens A Andersson

Waveform Pattern Recognition – A New Semiquantitative Method for Analysis of Fetal Aortic and Umbilical Artery Blood Flow Velocity Recorded by Doppler Ultrasound

A semiquantitative computerized waveform pattern recognition system for analysis of the fetal descending aortic and umbilical artery Doppler flow velocity waveforms is presented. Based on empirically and manually selected clinical recordings from both vessels, 11 computerized and normalized standard curves for the aorta (type curves A to K), and 10 curves for the umbilical artery (type curves a to j) were constructed. The best match between the normalized waveform and the standard curve was based on either the degree of absent diastolic flow or, in cases with positive diastolic flow, on the calculation of the least square sum of the difference. The pattern recognition was tested against conventional waveform indices and our older semiquantitative Blood Flow Class (BFC) system in 472 clinical consecutive Doppler recordings. A good correlation was found. This new relatively simple computer-based method for waveform analysis is now prospectively applied in clinical studies.

User profiling for pre-fetching or caching in a catch-up TV network

We investigate the potential of different pre-fetching and/or caching strategies for different user behaviour with respect to surfing or browsing in a catch-up-TV network. To this end we identify accounts and channels associated with strong or weak surfing or browsing respectively and study the distributions of hold times for the different types of behaviour. Finally we present results from a request prediction model and a caching simulation for the different types of behaviour and find that the results are relatively similar.

Traffic Shaping and Dimensioning of an External Overload Controller in Service Architectures

This paper investigates the dimensioning of a server used for external overload control in service architecture. Great savings can be obtained by an operator if the dimensioning analysis is performed correctly. As one of the main parts of this paper it is shown that Poissonian arrivals is a good assumption for some services in service architectures. Methods that can be used for dimensioning are presented and examples are provided

Analytic model for cross-layer dependencies in VDSL2 access networks

Recent changes in user employment of Internet based services, new deployment technologies for mobile networks as well as an ongoing realisation of fixed and mobile converged networks e.g. the EU FP7 project COMBO, are significant examples of enablers for increasing demands on DSL links. Investigating cross-layer dependencies between all layers in the OSI reference model becomes increasingly important. In this paper we present an analytical model and experimental results for the relation between impulse noise on a VDSL2 link and the effect this have on the network layer packet loss. We show how the packet loss rate is dependent not only on the disturbance signal level and periodicity but also on the link utilisation.