|
Current -Fractal Geometry Analysis of Time Series for Fault Identification, by Aysun Kaya.
Concluded: -Data-driven Self Organized Modeling using Reinforcement Learning in Autonomous Units, by Murat Karakurt. ABSTRACT The control of dynamic structures may not be always achieved by use of traditional control schemes. Since analysis, simulation and realization of these control policies almost always require an explicit "fairly-trustworthy" formulation of the phenomena, after which a satisfactory performance is expected. Yet, environmental uncertainty and nonlinear system behavior, which are ubiquitous in any and every system, necessitate a more complex, ever changing, adapting modelers. In this study, such an empirical modeler is being devised in the form of a self-organizing, self-tuning network. Thus a heuristic modeling and an intelligent reasoning are obtained based on sensing of the relevant physical qualities, fusing of the sensory information, if necessary, storage of data as either parameter adaptation or database through an optimal, intelligent paradigm. Optimal control policies are to be taught to the system by Reinforcement Learning, where an agent tries to discover the underlying rules within the environment through interactions. Supported by a 3,000 $ University research grant -Probabilistic Localization & Path Planning in Dynamic Environments, by Derya Gümüş. ABSTRACT Collecting data for an autonomous agent so as to gather information about its environment for mapping and path planning purposes is quite a complicated issue in static surroundings. This problem is further aggravated when the scene is cluttered with dynamic factors such as humans which lead to imprecise readings from sensors. To make things worse, many sensors have a limited range that eventually may cause the agent to get lost in vast open spaces. Fortunately, there are computational tools and techniques that "mitigate" the complexity of the problem. These run in different lanes, for example, dead reckoning is one way to facilitate the agent navigation, whereas entropy is a way of expressing certainty about its location. Here, we are trying to integrate intelligent tools in the navigation process, and to mimic the related human traits on the agent. We also believe nature is presenting us practical solutions, and by observing them, we may get hints in the analysis of the problem. Supported by a 2,000 $ University research grant -Control of a Manipulator Using Cerebellar Model Articulation Control (CMAC), by Murat Darka. ABSTRACT The Cerebellar Model Articulation Controller (CMAC) is a Neural Network topology commonly used in the field of robotic control which.was developed in the '70s by Albus. Robot motion requires complicated, explicit and an accurate model in the case of conventional control techniques. An accurate model is hard to implement due to uncertainties, thus an ad hoc controller, CMAC, seem to be a good choice. In this study, a robotic manipulator is intended to be controlled by a CMAC. As is the case for many CMAC architectures, these controllers are often plagued with instability that significantly deteriorates the network performance. A measure of the system performance is to be formulated so as to observe and then possibly control the system. Various performance measures are currently being studied. Following the analyses, a computer simulation along with physical experiments are to be conducted.
Courses: The lab currently offers 3 graduate level courses. ME 570 Computational Intelligence ME 578 Probabilistic Reasoning ME 588 Industrial Fault Detection and Identification
Aysun Kaya, Previously: Murat Karakurt, Murat Darka, Gürkan Bingül, Derya Gümüş.
Having been established only in 2002, the AID lab is a fledgeling new member among the others in the campus. Manned with a handful graduate students working in the intelligence related fields, the lab set his eyes in the relatively young field of AI, as well as well formulated fields of control and robotics. It is our intention to further our knowledge in the so-called areas, disseminate it and possibly establish close ties with the industry.
-Ozdemir S., "Measures of Uncertainty in Power Split Systems", submitted to the J of Mechanism & Machine Theory, 2004. -Bilir L., Alpay S., Ozdemir S., "Improved Responsive Perturbation with Trait-based Heterogeneous Populations", submitted to the J of Mathematics & Computers in Simulation, 2004. -Ozdemir S., "Power Transmission Entropy", submitted to the ASME J of Mechanical Design, 2004. -Ozdemir S., Darka M. "Forward Displacement Problem of a 6 DOF Parallel Manipulator by CMACs", submitted to the J of Applied Intelligence, 2004. -Alpay S., Bilir L., Ozdemir S., Ozerdem B. "Wind Speed Time Series Characterization by Hilbert Transform", submitted to the J of Renewable Energy, Feb 2004. -Ozdemir S., "Local Sparse Coding Control of CVPSTs", accepted for publication in the Int. J. of Vehicle Design, 2004. -Atan E., Ozdemir S. "Intelligent Modeling of the Transient Asperity Temperatures in Meshing Spur Gears" accepted for publication in the Journal of Mechanism and Machine Theory, 2004. -Ozdemir S., Schueller, J. "A New Hybrid CVT Design : CVPSTs", Heavy Vehicle Systems, A Series of the Int. J. of Vehicle Design, Vol. 9, No. 4, pp. 319–332, 2002.
Conference Papers : -Bozokalfa G., Ozdemir S., Atan E. "Langevin Equations for Stochastic Data Analysis in Noisy Time Series", To be presented at 4th Int Symposium on Intelligent Manufacturing Systems, Sept 6-8, 2004, Sakarya. -Bilir L., Alpay S., Ozdemir S.,"Study of Population of Heterogeneous Individuals", To be presented at 4th Int Symposium on Intelligent Manufacturing Systems, Sept 6-8, 2004, Sakarya. -Alpay S., Bilir L., Ozdemir S., Özerdem B. " Estimation of Wind Speed in Urla District Via Hilbert Transforms", CMES-04, July 14-16, 2004, Cappadocia, Turkey. -Atan E., Ozdemir S.,"Modeling of the Transient Asperity Temperatures in Meshing Spur Gears by Intelligence Methods". CMES-04, July 14-16, 2004, Cappadocia, Turkey. -Ozdemir S., Karakurt M., "A Combined SA-GA Optimization Variant for Networks". Current Advances in MD&P VII. p1167-1172, ISSN: 1366-0950, Jan 2004, Giza, Egypt.
Full-time graduate students, who can support themselves throughout their MS or PhD programs, are needed. The lab unfortunately cannot support individuals currently but the department, every now and then, grants assistanships to successful students. Please contact the Lab for more information if you would like to conduct research with us.
Fuzzy Logic
Neural Networks
Automatic Control
General AI
Other Links
|