SCALDIS Ltd was established in April 2010, by two scientists of Kazan State University: Raoul Nigmatullin and Sergey Osokin. Now, we are a united company consisting of professional scientists, programmers and engineers, which provide a variety of services in research & development and software development on all levels of difficulty. The spheres of our activities include researches in the field of dielectric and noise fluctuation spectroscopy, electron paramagnetic resonance and nuclear magnetic resonance, development of quality control and process control systems, as well as other software.
Our main goal - to create unique and innovative solutions in process control and monitoring of complex systems.
Scientific Basis
Group of our scientists led by Raoul Nigmatullin developed and tested three new original methods of signals and noise analyzing. These methods allow to represent arbitrary random series in some finite basis of small number of parameters without any model (usually Gauss) concepts about noise nature.
1. The first method is based on the procedure of optimal linear smoothing (POLS). The use of this method can optimally smooth out different data using the criterion of the relative error minimum. This approach was successfully tested on a great number of different data obtained from seismology, metrology, finances and infra-red spectroscopy.
2. The second method is based on recognizing the distribution function of the detrended noises of the strongly correlated systems. It is possible to show that there is also a distribution function for strongly correlated systems which approximately coincides with well known function of b-distribution. This means that all noises of strongly correlated systems can be compared to each other by the parameters of this distribution function.
3. The third method allows to expand the wide class of strongly correlated random functions to segment of Prony series and represents them in the form of amplitude-frequency characteristic. This method allows to compare the noises of different nature by their calculated amplitude-frequency characteristics and forms the base of noninvasive noise metrology.
Suggested original methods called NIMRAD (Non Invasive Methods of Reduced Analysis of Data) can get very wide use for detection of superweak signals, random overshoots, particularly to control the work of complex quantum systems, to control the quality of the production and for monitoring the systems with complex organisation (electromotors, automobile engine, aircraft engine, electronic modules of transducers and sensors, living organisms), in criminalistics for detection of the explosive materials, drugs and e.t.c.).
Nowadays for noise analysis two basic methods are used: Fourier analysis and Wavelet analysis, which have several important disadvantages.
In the case of Fourier analysis one of the most obvious problems is the initial assumption about the Gaussian nature of noise. In fact this assumption is incorrect and causes mistakes in further calculations. The next disadvantage is the invasiveness of this method. So, if the researcher uses the inverse conversions to results obtained by using Fourier conversions, he receives data different from the input data.
Wavelet analysis solves some of these problems, but it has its own disadvantages. The choice of wavelet (function which helps to choose basis for the further calculations in smaller number of parameters) is made by expert. So, using this method, we lose ability to automate the process of analyzing noise and signals.
NIMRAD methods developed by our team works with a different nature of noises, and use non invasive calculations. Our methods were tested for different physical, biological and chemical processes. Nowadays we develope several software products that implement our methods for a variety of devices.