[en] Three basic methods that are extensively applied at JINR to process recent experimental data are reviewed, namely, robust methods of mathematical statistics, artificial neural networks and wavelet analysis. This review primarily covers studies in which scientists from the Laboratory of Information Technologies participated, in particular, in collaborations with the leading centers of physics, such as CERN, DESY, BNL, GSI, etc. The main principles of the reviewed methods and the most useful and promising examples of their applications are discussed

[en] New structure and training algorithms of the RBF-type neural network are proposed. An extra neuron layer is added to realize the principal component method. A novel training algorithm is designed for training each separate neuron in the hidden layer, which guarantees the efficiency and finiteness of the training procedure. Results were obtained for a variety of problems. In particular, the results of human face recognition look very promising

No abstract available

[en] The modification of the deformable template method (DTM) application to the problem of track finding and track parameter determination for data detected with high pressure drift tubes (HPDT) in the design of ATLAS for the muon spectrometer experiment is proposed. Our DTM applications consist of two parts, according to two stages of the study. The first part relates to the stage of HPDT study on the CERN muon beam (BEAM-TEST) with the simplest one-prong events without noise signals, where the main obstacle is the left-right ambiguities for each tube. In the second part more complicated HPDT data are to be handled with noise signals. It was shown that the suggested DTM development solves the problem of track recognition and track parameter determination for both noiseless and noise data. Results are obtained on the real beam test data and on data simulating the muon spectrometer on the basis of HPDT. 14 refs., 10 figs

[en] An algorithm of track recognition in an uniform magnetic field is proposed for the drift straw tube detecting system of solenoidal geometry. The problem solution is given for (x,y) plane perpendicular to the magnetic field. Our algorithm is elaborated on the basis of (1) sequential histogramming method, which is, in fact, a modification of the Hough transform and (2) a modification of the deformable template method followed by a special procedure of the parameter correction. Being tested on simulated events our method shows satisfactory efficiency and accuracy in determination of particle momenta. 9 refs., 2 figs

[en] Problems of direct processing of raw Ring Imaging Cherenkov (RICH) data for the high granularity detectors of the CERES type are discussed. The robust technique is applied to ring fitting. It is based on a least-square method with recalculated weights depending not only on the distance of pads from the fitted circle but also on the signal amplitudes in pads. The proposed method leads to reliable parameter reconstruction from raw data for a single ring, as well as for two overlapping rings. Computed ring parameters are then successfully used for the particle identification

[en] The study of the functioning of detectors like a multiwire proportional chamber has led to the necessity of taking into account its discrete structure and of prefiltering data before processing. The filtering method for tracks in this discrete detector based on the cellular automaton is described. Results of the application of this method to experimental data (the spectrometer ARES) are quite successful: A high degree of cleaning from noise and the restoring of missing clusters give three-fold reduction of raw data with the partial track recognition. The percentage of good events is increased and their further processing is considerably simplified and accelerated. The described cellular automaton for track filtering is suitable for parallel computing and could be also used in the on-line mode. (orig.)

[en] In this paper, results of applying a method of internal alignment of HERA-B OTR PC chambers are discussed. The method is based on simultaneous fit of the track and alignment parameters using Millepede matrix reduction and singular value decomposition. After extensive tests of the method on different Monte-Carlo models, a number of studies have been done using real data taken by HERA-B in a 2002-2003 run period. Misalignment influence on reconstruction of individual tracks and physics signals, decays of J/ψ and K_S⁰, has been studied

[en] In this paper a method of internal alignment of HERA-B OTR PC chambers is discussed. The method is based on simultaneous fit of the track and alignment parameters using Millepede matrix reduction and singular value decomposition. Software which implements this idea has been developed, the method has been studied on Monte Carlo models with different levels of simulation. A method generalization for the case of track nonlinear model has been proposed

[en] Two algorithms of ring recognition, a standalone ring finder (using only RICH information) and an algorithm based on the information from vertex tracks are described. The fake ring problem and its solution using a set of two-dimensional cuts or an artificial neural network are discussed. Results of a comparative study are given. All developed algorithms were tested on large statistics of simulated events and were then included into the CBM framework for common use