Steffen HARNEIT

  • Citations Per Year
Learn More
Self-organizing feature maps and feedforward nets trained with the backpropagation algorithm are the most common used and the most analyzed artificial neural nets. Although a combination of them, called counterpropagation, already had been proposed, its full capabilities had not been recognized. In this paper an extension of this net structure is proposed,(More)
Hybrid classification structures of buried object in different soils using time domain metal detectors are presented in this paper. This approach consists of self organising memories and backpropagation nets combined with the inter-neural whu-structures, whereby special pre-processing methods like DLS and fitting-methods are used to extract the feature(More)
It is reported by experienced deminers that the existence of moisture in soils complicates the detection of buried land mines. In this work we examine the influence of water in two different sample soil types (sand and clay soil) on the magnetic field induced by a mine surrogate and thus on the quality of the detector's audible signal when using a frequency(More)
Originally, mines were detectable because of their more or less high metal content. Unfortunately, in the fast developing mine technologies, new materials that may last for many years, have made it possible to produce varieties of mines that are practically undetectable with the existing methods because of very low metal contents. As we showed, this fact(More)
In this work, a new kind of autonomous solar still is proposed. This construction is used in the arid countries with intense solar irradiation and has shortage of potable water. The goal was to concept and to build a physical and mathematical model of solar sea water distiller. The model design and the comparative simulations with the measurement results(More)
In this work we propose an end-user supporting system for humanitarian demining tasks to semi-automatically classify signals of time domain metal detectors. Our multi-stage system consists of a first module to smooth the raw signals, followed by a neural feedforward net to classify the received signal's decay curve of the localized object at each sensor(More)
  • 1