 |
View: 1-15 | 16-30 | 31-45 | 46-60 | 61-69
II. Intelligent Control Systems
Tuning of PI Speed Controller in DTC of Induction
Motor Based on Genetic Algorithms and Fuzzy Logic Schemes
S.M. Gadoue, D. Giaouris and J. W. Finch
Pages: 85-90
Abstract: PID controllers are very common in industrial systems applications. The tuning of these controllers is governed by system nonlinearities and continuous parameter variations. In this paper, a complete and rigorous comparison is made between two tuning algorithms. The PI controller was used in a speed control loop in a Direct Torque Control (DTC) scheme applied on an induction motor. The first method applies off-line genetic algorithm (GA) strategies and the other one makes use of an online Fuzzy Logic (FL) tuning scheme. DTC is then tested with the two schemes for two cases, with normal operating conditions and with a sudden change in load torque applied to the motor. Results obtained show that the fuzzy logic on-line tuning technique can provide better speed control performance when system parameter variations occur. On the other hand for nominal operation the genetic algorithms scheme is preferred.
Intelligent Technologies
for Energy Efficiency and Comfort in a Building Environment
A.I. Dounis and C. Caraiscos
Technol. Educ. Inst. of
Pages: 91-95
Abstract: In this work we present a multi-agent control system that can be applied to a building environment in order to obtain energy conservation and occupants’ comfort. The system consists of a number of local controllers-agents that are coordinated by an intelligent supervisor. The proposed architecture is based on the concept of intelligent agents that has been introduced recently in the area of artificial intelligence.
Intelligent Monitoring
of Robotic Systems with PIC Microcontrollers and a Petri-Net based Approach
D. Koutandos
McQuay Hellas S.A.,
Pages: 96-101
Abstract: This paper investigates the development of an intelligent and low-cost monitoring system for simple robotic systems, considering the need of modern industries for fault diagnosis and identification for early detection of failures for maintenance and managerial activities. Grippers are investigated as part of a pick-and-place system and a PIC microcontroller-based monitoring system is developed. The use of Petri-nets becomes a vital part of this work and allows the modeling of the process. A distributed monitoring system employing industrial networks and internet technology, developed by the Intelligent Process Monitoring and Management (IPMM) Centre group, is used to monitor the operation of an industrial pneumatic gripper. Overall, the project reveals the efficacy of a low-cost, microcontrollerbased monitoring system, using a Petri-net approach, for robotic operations and by doing so it demonstrates its important advantages in intelligence and flexibility.
Intelligent
Robust-Switching PID Controller Design for a Micro-Actuator
M. Vagia, G. Nikolakopoulos and A. Tzes
Pages: 102-107
Abstract: In this article the design of an intelligent robust controller for a Micro–Actuator (µ−A) is presented. The µ−A is composed of a micro–capacitor, whose one plate is clamped while its other flexible plate’s motion is constrained by hinges acting as a combination of springs and dashpots. The distance of the plates is varied by the applied voltage between them. The dynamics of the plate’s rigid-body motion results in an unstable, nonlinear system. The control scheme is applied to multiple linear time invariant models of the µ − A from the linearization process in multiple points. This control scheme is constructed from: a) a feedforward controller which stabilizes the micro–actuator around its nominal operating point, b) a robust-switching PID controller which gains are firstly tuned via the utilization of Linear Matrix Inequalities (LMIs) and secondly they are switched during the simulation depending on the displacement of the upper plate, and c) an intelligent prefilter which shapes appropriately the reference signal. The resulting overall control scheme is applied to the non–linear model of the µ − A where simulation results are presented to prove the efficacy of the suggested scheme.
Application of GA Based Fuzzy Neural Networks for Measuring Fouling in Condenser
F. Shao-Sheng1 and W. Yao-Nan2
1
Pages: 108-112
Abstract: A novel approach for online measurement of fouling in condenser is proposed in this paper. In the approach, terminal temperature difference is chosen to reflect fouling state, B-spline membership fuzzy neural network is employed to approximate off-design condition terminal temperature difference, which separates the influence imposed by fouling on terminal temperature difference from other factors. Since the selection of the weighting factors, the knot positions and the control points of the B-spline membership fuzzy-neural networks is crucial to obtaining good approximation for complex nonlinear systems, a genetic algorithm with an efficient search strategy is developed to optimize these variables. Based on the approach, an experimental system is developed and experiment on an actual condenser is carried out. The results show the approach measures the fouling correctly, and is more effective than thermal resistance method or heat transfer coefficient method under the condition of blocked tubes or excessive amount of air in condenser.
Prioritized Adaptive
Model Predictive Control Using Evolutionary Algorithms
E. Aggelogiannaki and H. Sarimveis
National Technical
Pages: 113-118
Abstract: Adaptive Model Predictive Control (MPC) configurations are quite popular methodologies for successful control of dynamic time varying systems. Most adaptive schemes include the persistent excitation requirement as an additional hard constraint of the optimization problem. In this paper an alternative approach is attempted, by using the principles of multiobjective optimization. A prioritized optimization problem is formulated, considering the persistent excitation as the top priority objective. Afterwards, an objective function is assigned to each one of the remaining control goals. This way, the adaptive capabilities of the methodology are exploited and the time consuming tuning procedure to weigh the different control goals in one objective function, is avoided. An additional innovation of the proposed configuration is the utilization of an improved evolutionary algorithm in order to meet the complexity and non convexity introduced by the persistent excitation requirement. The overall proposed configuration is evaluated through the application to a continuous stirred tank reactor. The produced results are superior compared to the performance of a conventional MPC scheme.
A Fuzzy Rule Based Approach for Storing the
Knowledge Acquired from Dynamical FCMs
Y. Boutalis1, T. Kottas1, B. Mertzios2
and M. Christodoulou3
1
Pages: 119-124
Abstruct: Fuzzy Cognitive Maps (FCMs) have found many applications in social - financial - political problems. In this paper we propose a method of FCM operation, which can be used to represent and control any real system, including traditional electro-mechanical systems. In the proposed approach the FCM reaches its equilibrium point using direct feedback from the node values of the real system and the limitations imposed by the control objectives for the node values of the system. To avoid intensive interference of the updating mechanism with the real system, a technique is proposed which stores the previously encountered operational situations in fuzzy if-then rule database. The proposed methodology is tested by simulating the operation of a hydroelectric plant.
A Novel Data Fusion Approach in an
Integrated GPS/INS System Using Adaptive Fuzzy Particle Filter
A. Asadian1 and B. Moshiri1 and A.
Khaki-Sedigh2
1Univ. of
Pages: 125-130
Abstract: In this paper we propose a new data fusion method based on particle filtering and fuzzy logic in order to adaptively integrate global positioning system and strapdown inertial navigation system (GPS/SDINS). This approach will reduce the dependence of the stable solution on stochastic properties of the system which is a function of vehicle dynamics and environmental conditions So the proposed scheme will enhance the estimation performance in comparison with generic particle filter specially in the case of facing modeling uncertainty. It will also give us more reliable solution when encountering satellite signal blockage as a probable problem in land navigation. The results have clearly demonstrated that the hybrid fuzzy particle filter would improve the guidance from the point of accuracy and robustness to the mentioned problems.
Solving University
Timetabling Problems Using Advanced Genetic Algorithms
S. Kazarlis1, V. Petridis2 and P.
Fragkou2
1Technol. Educ. Inst. of
Pages: 131-136
Abstract: Timetabling problems together with scheduling ones constitute a class of difficult to solve combinatorial optimization problems that lack analytical solution methods. As such, these problems have attracted researchers from a number of disciplines, like Operations Research and Artificial Intelligence, who have proposed a number of methods for solving them. In this paper we present a method based on Genetic Algorithms (GAs), to solve university course timetabling problems. This method incorporates GAs using an indirect representation based on event priorities, Micro-GAs and heuristic local search operators in order to tackle a real world timetabling problem. The problem on which the method is applied and tested is a real case and comes from a Technological Educational Institute of Greece. The GA solution is compared to the manmade one produced by the institute’s staff and the comparative results are discussed.
Neural Network for Fault Detection and
Isolation of the Three-Tank System
P. Tzionas, D. Krontsos and S. Papadopoulou
Alexander Technol. Educ.
Inst. of
Pages: 137-142
Abstract: This paper presents the design, training, verification and validation of a neural network architecture capable of early fault detection and fault isolation in a typical three –tank system. Certain fault types are induced to the system and its behavior is monitored. Parameters such as water-level and temperature in the tanks, together with delayed samples are used to design, train and validate the neural network architecture. The neural network is further tested on a set of signal values derived from subsequent operation of the system, with considerable success.
A Comparison and
Application of Pattern Recognition Techniques in Materials Selection Using Dimensionless Numbers
J.W. Schmidt-Castaneda, F. J. Sandoval and D. J. Dorantes- Gonzalez
Inst. Tecnol. y de Estudios Superiores de Monterrey, Mexico
Pages: 143-147
Abstract: The objective of this paper is to present a tool and a comparison of techniques that can help in the material selection process based on pattern recognition and using dimensionless parameters from Ð theorem. It is known that the main disadvantage in materials selection is that a huge database is needed to make the selection and also to make clusters of materials with some characteristics, so we need an easier and automatic way to cluster materials with respect to their physical and mechanical properties, and particularly for material machinability. This paper also shows a possible application related to material selection using this approach to complement the conceptual design of a part coding system using neural network techniques.
III. Robotics
A Controller for
Changing the Yaw Direction of an Underactuated Unicycle Robot
S. Majima and T. Kasai
Pages: 148-153
Abstract: Several researchers have developed unicycle robots and associated control systems. We have also developed a unicycle robot that consists of a body, a disk driven by a dc motor for rotation, and a wheel driven by another dc motor to move the body. Because the disk is rotated in the frontal plane, there is no direct force to change the yaw direction of the robot. Hence conventional controllers cannot change the yaw direction. The present paper proposes a control method for changing the yaw direction by using the interaction between individual motions in the yaw, pitch, and roll directions. Simulation and experimental results show that the control method can control the yaw direction of the unicycle robot while stabilizing its posture. An observer to estimate the robot posture is also developed.
Modelling and Interfacing Remote Virtual
Robots
Z. Doulgeri, N. Zikos and A. Delopoulos
Aristotle Univ. of
Pages: 154-159
Abstract: A formal generic model for simulating the kinematic behavior of virtual robotic arms is presented in this work. We introduce an algorithmic procedure, called Virtual Robot Simulation (VRS) Engine, for updating the trajectory of the virtual robot when motion commands are presented for execution. The trajectory itself is modelled as a list consisting of time intervals where robot joints follow a piecewise polynomial path. In addition, the same engine is equipped with the functionality of responding to position requests at arbitrary time instances. Constraints imposed by the operation of VRS in real-time mode are also explored. User interfaces to the virtual robot - including remote access through data networks - are also modelled and the effects of possible communication delays are explored. The combination of the proposed simulation environment with a synchronized real robot is proposed as a means to overcome the lack of direct visual feedback in tele-robotic applications.
Controlling the
Loop-Gain for Robust Adaptive Control of a Mechatronic System
C. Westermaier, H. Schuster and D. Schroeder
Technical
Pages: 160-165
Abstract: This paper concerns the question of applicability of adaptive control strategies in real environments. Because of unrobustness to unmodeled dynamics - especially dead time - model reference adaptive control with all its positive features can not be implemented in industry. But it can be shown that an additional gain-controller within the MRAC-concept leads to a robust adaptive controller applicable to real systems. In this context, the paper gives a possibility of closing the gap between theory and praxis in the field of adaptive control. As a case study, a two-mass flexible servo system with unknown inertia, spring and damping constants is investigated while the dynamics of the power converter, speed-sensor and further unknown and time-varying dead-times can be neglected. The goal is a perfect dynamic tracking of the load-mass speed with a smooth control output.
Optimal Kinodynamic
Planning for Autonomous Vehicles
S. Vougioukas
Aristotle Univ. of
Pages: 166-171
Abstract: This paper presents a two-phase kinodynamic planning algorithm, which uses a randomized planner to compute low-cost robotic motions, and optimal control to locally optimize them. The direct transcription approach is used with an NLP numerical optimization algorithm. The algorithms are tested on motion planning for a non-holonomic autonomous vehicle. The results indicate that the two-phase approach is effective in computing optimal motions. However, the dense algebra NLP solver is very time consuming for the optimization of long paths and sparse algebra solvers should be utilized.
View: 1-15 | 16-30 | 31-45 | 46-60 | 61-69 | Back to top