Gheorghe PUSCASU
Assistant Professor
PhD - "Contribution to Learning Machines using Neural Networks and Methods
of Pattern Recognition for
Biotechnological Process Control", "Dunarea de Jos" University of
Galati, Romania, 1996.
BEng - Faculty of Automatic Control and Computer Science, Politehnical
University of Iassy, Romania, 1987.
Email:
Gheorghe.Puscasu@ugal.ro
Telephone:
+40 36 46 01 82
Fax:
+40 36 46 01 82
Office:
Room 503, Y
Teaching
System Identification and System Control, Artificial Neural
Networks and Fuzzy Techniques in Nonlinear Process Control, System Theory,
Data Acquisition and Signal Processing, Image Processing and Pattern Recognition.
Research
- Development of new learning algorithms and architectures of neural
networks;
- Applications of the neural networks in pattern recognition and control;
- Applications of neural networks to 1D, 2D signal analysis and processing;
- Fault detection using pattern recognition techniques;
- Neural techniques used in system identification;
- Development of two large applications based on neural networks for
criminalistic expertise and the recognition of a large vocabulary;
- Development and implementation of some equipments for process control
used in teaching activities.
Activities
1997-1998: Associate Professor at the Department of Automatic Control,
“Dunarea de Jos” University of Galati, Romania.
1994-1997: Senior Lecturer at the Department of Automatic Control,
“Dunarea de Jos” University of Galati, Romania.
1990-1994: Lecturer at the Department of Automatic Control, “Dunarea
de Jos” University of Galati, Romania.
1987-1990: Software Engineer at the Computer Centre of the Romanian
Railway Department of Galati, Romania.
Publications
Gh. Puscasu - Sistems Identification, practical course, University
of Galati, 1997
Gh. Puscasu, S. Caraman, V. Palade - Laboratory Guide in Process
Control by Systems and Equipment with Microprocessor, “Dunarea de Jos”
University of Galati, 1993
Gh. Puscasu, C. Milos - Vowels Recognition Using Neural Networks,
the Annals of “Dunarea de Jos” University of Galati, vol. III, pp.
37-40, 1994
Gh. Puscasu, E. Ceanga, C. Milos - Capabilities of multi-layer neural
networks concerning elimination of different types of neurons, the
Annals of “Dunarea de Jos” University of Galati, vol. IV, pp. 23-29,
1995
Gh. Puscasu, E. Ceanga, V. Palade - Connections between RBF
neural networks, Kohonen algorithms and fuzzy sets, the Annals
of “Dunarea de Jos” University of Galati, vol. VI, 1997
Gh. Puscasu, L. Nica - DC Motor's Speed Control Using Neural Networks,
Electromotion'95 Symposium, Cluj-Napoca, Romania, pp. 44-47, 1995
Gh. Puscasu, L. Nica - Evaluation Criteria for Kohonen Neural
Network Results. New Applications of Kohonen Neural Networks,
5th Symposium on Automatic Control and Computer Science, Iassy-Romania
, vol. II, pp. 279-282, 1995
Gh. Puscasu - The DC Motor’s Identification and Control Using Neural
Networks, OPTIM '96, Brasov-Romania, pp. 1707-1714, 1996
Gh. Puscasu, C. Milos - A New Approach for Multi-layer Neural Networks,
the International Conference on Technical Informatics, CONTI’96,
Timisoara-Romania, pp. 41-48, 1996
Gh. Puscasu - The Non-linear Process Control
with Internal Model and Multi-layer Neural Networks, Control
of Oscillations and Chaos, St. Petersburg-Russia, 1997
This information provided by G.Puscasu; July
2000.