[show abstract][hide abstract] ABSTRACT: The sensitivity of a Cherenkov imaging telescope is strongly dependent on the
rejection of the cosmic-ray background events. The methods which have been used
to achieve the segregation between the gamma-rays from the source and the
background cosmic-rays, include methods like Supercuts/Dynamic Supercuts,
Maximum likelihood classifier, Kernel methods, Fractals, Wavelets and random
forest. While the segregation potential of the neural network classifier has
been investigated in the past with modest results, the main purpose of this
paper is to study the gamma / hadron segregation potential of various ANN
algorithms, some of which are supposed to be more powerful in terms of better
convergence and lower error compared to the commonly used Backpropagation
algorithm. The results obtained suggest that Levenberg-Marquardt method
outperforms all other methods in the ANN domain. Applying this ANN algorithm to
$\sim$ 101.44 h of Crab Nebula data collected by the TACTIC telescope, during
Nov. 10, 2005 - Jan. 30, 2006, yields an excess of $\sim$ (1141$\pm$106) with a
statistical significance of $\sim$ 11.07$\sigma$, as against an excess of
$\sim$ (928$\pm$100) with a statistical significance of $\sim$ 9.40$\sigma$
obtained with Dynamic Supercuts selection methodology. The main advantage
accruing from the ANN methodology is that it is more effective at higher
energies and this has allowed us to re-determine the Crab Nebula energy
spectrum in the energy range $\sim$ 1-24 TeV.
Nuclear Instruments and Methods in Physics Research Section A Accelerators Spectrometers Detectors and Associated Equipment 01/2013; 708. · 1.14 Impact Factor
[show abstract][hide abstract] ABSTRACT: We report an inter-comparison of some popular algorithms within the artificial neural network domain (viz., local search algorithms,
global search algorithms, higher-order algorithms and the hybrid algorithms) by applying them to the standard benchmarking
problems like the IRIS data, XOR/N-bit parity and two-spiral problems. Apart from giving a brief description of these algorithms,
the results obtained for the above benchmark problems are presented in the paper. The results suggest that while Levenberg-Marquardt
algorithm yields the lowest RMS error for the N-bit parity and the two-spiral problems, higher-order neuron algorithm gives
the best results for the IRIS data problem. The best results for the XOR problem are obtained with the neuro-fuzzy algorithm.
The above algorithms were also applied for solving several regression problems such as cos(x) and a few special functions like the gamma function, the complimentary error function and the upper tail cumulative x
2-distribution function. The results of these regression problems indicate that, among all the ANN algorithms used in the present
study, Levenberg-Marquardt algorithm yields the best results. Keeping in view the highly nonlinear behaviour and the wide
dynamic range of these functions, it is suggested that these functions can also be considered as standard benchmark problems
for function approximation using artificial neural networks.
KeywordsArtificial neural network-benchmark problems-function approximation-special functions
[show abstract][hide abstract] ABSTRACT: An Artificial Neural Network-based error compensation method is proposed for improving the accuracy of resolver-based 16-bit encoders by compensating for their respective systematic error profiles. The error compensation procedure, for a particular encoder, involves obtaining its error profile by calibrating it on a precision rotary table, training the neural network by using a part of this data and then determining the corrected encoder angle by subtracting the ANN-predicted error from the measured value of the encoder angle. Since it is not guaranteed that all the resolvers will have exactly similar error profiles because of the inherent differences in their construction on a micro scale, the ANN has been trained on one error profile at a time and the corresponding weight file is then used only for compensating the systematic error of this particular encoder. The systematic nature of the error profile for each of the encoders has also been validated by repeated calibration of the encoders over a period of time and it was found that the error profiles of a particular encoder recorded at different epochs show near reproducible behavior. The ANN-based error compensation procedure has been implemented for 4 encoders by training the ANN with their respective error profiles and the results indicate that the accuracy of encoders can be improved by nearly an order of magnitude from quoted values of ~6 arc-min to ~0.65 arc-min when their corresponding ANN-generated weight files are used for determining the corrected encoder angle. Comment: 16 pages, 4 figures. Accepted for Publication in Measurement Science and Technology (MST)
Measurement Science and Technology 11/2009; · 1.44 Impact Factor
[show abstract][hide abstract] ABSTRACT: The energy estimation procedures employed by different groups, for determining the energy of the primary γ-ray using a single atmospheric Cherenkov imaging telescope, include methods like polynomial fitting in SIZE and DISTANCE, general least square fitting and look-up table based interpolation. A novel energy reconstruction procedure, based on the utilization of Artificial Neural Network (ANN), has been developed for the TACTIC telescope. The procedure uses a 3:30:1 ANN configuration with resilient backpropagation algorithm to estimate the energy of a γ-ray like event on the basis of its image SIZE, DISTANCE and zenith angle. The new ANN-based energy reconstruction method, apart from yielding an energy resolution of ∼26%, which is comparable to that of other single imaging telescopes, has the added advantage that it considers zenith angle dependence as well. Details of the ANN-based energy estimation procedure along with its comparative performance with other conventional energy reconstruction methods are presented in the paper and the results indicate that amongst all the methods considered in this work, ANN method yields the best results. The performance of the ANN-based energy reconstruction has also been validated by determining the energy spectrum of the Crab Nebula in the energy range 1–16 TeV, as measured by the TACTIC telescope.
Nuclear Instruments and Methods in Physics Research Section A Accelerators Spectrometers Detectors and Associated Equipment 04/2009; · 1.14 Impact Factor
[show abstract][hide abstract] ABSTRACT: The sensitivity of a Cherenkov imaging telescope, is strongly dependent on the rejection of the cosmic-ray background events. Some of the methods which have been used to achieve this segregation include methods like Supercuts, Maximum likelihood classifier, Kernel methods, Fractals Wavelets, Factorial Moments, Random Forest etc. While the segregation potential of neural network classifier has been investigated in the past with modest results, a detailed study using some recently incorporated popular algorithms in ANN (e.g. Conjugate Gradient methods, Radial Basis function algorithm, Simulated Annealing technique, Levenberg-Marquardt algorithm etc.) has not been done so far. The main purpose of this paper is to study the gamma / hadron segregation potential of these algorithms, by applying them to the Monte Carlo simulated data for the TACTIC imaging telescope. The results suggest that the algorithms based on Higer order neurons and Levenberg- Marquardt method are superior to the widely used Dynamic Supercuts procedure, for rejecting the unwanted hadronic background This paper was given the Best Poster Award at the 25 th meeting of the Astronomical Society of India,
[show abstract][hide abstract] ABSTRACT: A novel energy reconstruction procedure based on the utilization of Articial Neural Network has been devel- oped for the TACTIC atmospheric Cerenkov imaging telescope to estimate the energy of the primary gamma- rays in the TeV energy range. The procedure uses a 3:20:1 conguration of the ANN with resilient back- propagation training algorithm to estimate the energy of a -ray like event on the basis of its image SIZE, DISTANCE and zenith angle. The results obtained by using the CORSIKA code simulated data suggest the energy resolution of the telescope is 40 for retaining 90 of the -ray events in a particular energy bin which is comparable to the energy resolution of other single element imaging telescopes. Details of the energy estimation procedure along with results obtained by determining the Crab Nebula energy spectrum in the energy range 1-16 TeV as measured by the TACTIC telescope are presented in the paper. Estimating the energy of primary -rays on the basis of their Cerenkov light content on the ground is an important advantage which endows the atmospheric Cerenkov technique with calorimetric capability. Although the light intensity in an image (also known as image SIZE), recorded by a single atmospheric Cerenkov imaging telescope represents a key parameter for determining the energy of the primary -ray, one has to also consider its dependence on the core-position and zenith angle for attaining an energy resolution in the range of 30- 40 . Since, the core distance can not be measured directly with single telescopes, an approximate measure of this can be found by using the 'DISTANCE' parameter of the Cerenkov image. The -ray energy estimation thus becomes a function of three variables ( viz., SIZE, DISTANCE and zenith angle), whose analytical form is not known. Given the inherent power of ANN (Articial Neural Network) to effectively handle the multivariate data tting we have, in this paper, developed an ANN-based energy estimation procedure so that after detecting a candidate -ray source, one can also determine its energy spectrum. 2. Salient features of the TACTIC telescope The TACTIC (TeV Atmospheric Cerenkov Telescope with Imaging Camera) gamma-ray telescope has been set up at Mt. Abu ( 24.6 N, 72.7 E, 1300m asl), a hill resort in Western India, for studying emission of TeV -rays from celestial sources. The telescope deploys a F/1 type tracking light collector of 9.5 m area made up of 34 x 0.6 m diameter, front coated spherical glass facets which have been prealigned to produce an on-axis spot of 0.3 diameter at the focal plane. The telescope uses a 349-pixel, photomultiplier tube (ETL 9083UVB) based imaging camera with a uniform pixel resolution 0.3 and a eld of view 6 x6 to take a fast snapshot of the atmospheric Cerenkov events produced by an incoming cosmic ray particle or a gamma ray photon with an energy above 1 TeV. The back end signal processing hardware of the telescope is based on inhouse developed medium channel density NIM and CAMAC modules. The data acquisition and control
[show abstract][hide abstract] ABSTRACT: A novel image-cleaning method, based on the utilization of Artificial Neural Network (ANN) is shown here to `correctly' select more pixels in a Cerenkov image than in the conventional approach.
Bulletin of The Astronomical Society of India - BULL ASTRON SOC INDIA. 01/2003; 31:495-496.