Sharif Digital Repository

Control of parallel robots continues to be a challenging research topic. The structural variety of such robots dictated by the application and function makes such control strategies even more unique. As these robots often require nonlinear control schemes, such research has been focused towards two general controlling computational intelligent methods, namely adaptive (and optimal) control and artificial intelligence (e.g. fuzzy and Neural Net) based control. This study considers Neural Net (NNs) and fuzzy control method in order to propose a control scheme for a newly introduced 5-link parallel robot which has 3 degrees of freedom operating in a single plane environment (i.e. two... 

This thesis describes different neural networks models and an approximation based neural network controller for autonomous underwater vehicles (AUVs). The online multilayer perceptron neural networks (OMLPNN) have been designed to perform modeling of AUVs of which the dynamics are highly nonlinear and time varying. The online recurrent multilayer perceptron neural networks (ORMLPNN) have been additionally designed to generate a memory to pervious states and increase the performance of the modeling. The designed OMLPNN and ORMLPNN with the use of backpropagation learning algorithm have advantages and robustness to model the highly nonlinear functions. The proposed neural networks architecture... 

Stabilizing periodic orbit of fractional order chaotic system via linear state feedback is the subject of this research. Firstly, for detection of the unstable periodic orbits (UPO) in fractional order chaotic systems, the famous shooting method for integer order systems is extended to fractional order systems. After that, for stabilizing the fractional order system a hypothesis is stated and a theorem is proved in a specific condition. Besides the proven theorem, some examples are presented as evidences that verify the hypothesis. So, through the hypothesis, common methods for stabilizing integer order systems can be extended to the fractional order systems. Finally, for stabilizing the... 

Smart structures are the combination of structure, smart material, electronics and control technologies. Changing the shape of the structures by smart actuators is one of the most important applications of Shape Memory Alloy (SMA) in these structures. Having used these actuators, we can effortlessly and continuously deform and reshape the structures. Nevertheless, working with SMA actuated smart structures has one obvious drawback that is their hysteretic and nonlinear behavior, making modeling and control of these structures complex. Another difficulty in the control of smart structures is their great sensitivity to the actuating force that reduces the controllability of these structures.... 

Modeling and Control are fundamental issues for fish robots, which can play basic roles in design, optimization, fabrication, and eventually utilization of them. A review on the literature reveals the shortage of an analytical closed form model with little simplifications and high precision, and also control works based on such models. Studying LAEBT theory, it is shown that this theory is suitable for determining the forces produced due to the tail movements considered in the present work, and then it is used to determine the forces. Experimental investigations by means of the setup made for this purpose showed that the obtained equations for the forces have got acceptable precision.... 

Diabetes is a global problem as a silent epidemic disorder. Diabetes Mellitus has involved at least five percent of our population. It will cause so many complications and end-organs effects if it is not controlled. Today, treatment and control of diseases using computational intelligent systems has been paid a lot of attention to. That’s why this study has taken treatment and control of diabetes into consideration. So, the neural networks have been used to manage and treat diabetes. To do this, real data from diabetic patients has been gathered and based on the data, a model has been obtained to show the effect of insulin on the amount of blood glucose. Then, a neural network controller has... 

Control of two-arm flexible robot is studied in this research. The goal is to control the robot in order to carry a fluid tank installed in the top of the last arm, while the vibration of fluid must be damped. The equations governing the system are extracted by the energy method while Euler-Bernoulli beam is selected to model the elastic members. In order to develop the equations, all of the potential and kinematic energies due to rigid and elastic motions are calculated, then a proper controller is designed for the system to accomplish the considered missions. Since the system is classified as an under-actuated system, the exact feedback linearization method is not practical to control the... 

Polyethylene terephthalate (PET) is known as the most important polyester in the world. This polymer is the raw material for synthetic fibers, films and plastic objects and widely used in textiles and food industry. In this project transesterification and polycondensation reactors of a semi-batch process with dimethyl terephthalate/ethylene glycol feedare simulated. First simulations areperformed for reactors in the laboratory scale.Then the reactors are scaled up to industrial scale in order to produce 100 tons per year of PET product. The optimal temperature trajectory for the fisrt reactor is obtained based on a performance index. This performance index includes the reaction time and... 

One of the reducing and controlling tools of high speed crafts is intelligent interceptor. Interceptor is a small blade that is originally installed vertically in stern. This blade creates a high pressure area under hull of the craft and produces lift force and a moment to the bow. The amount of lift force and moment can be changed by controlling the height of the blade. The intelligent interceptor (with automatic change in height of the blade) does the same thing. In this study, structural and software design of intelligent interceptor is done. In software field, optimal and PID controller are designed for interceptor system and the performances are studied in MATLAB SimuLink. The results... 

An event must be modeled in a way that either reflects a comprehensive perspective for the event or acts in an especial part of that event, roperly. The main aim of this thesis is to control. Therefore the models will be accepted whenever the suggested mathematical models can move towards controller final target, desirably.This thesis proposes a linear mathematical model, five nonlinear models and a simple model based upon Convolution, Fuzzy Regression and Neural Networks techniques for Acquired Immune Deficiency Syndrome (AIDS), respectively. The proposed models were achieved through studying 300 HIV+ Patients who were under Highly Active Antiretroviral Therapy (HAART) approach in Iranian... 

Microrobots design and manufacturing has been one of interesting fields in robotics in recent decades. They have been designed in various types so far. In this paper a two-leg microrobot is simulated and controlled by using LuGre friction model.
The proposed microrobot is driven by a piezoelectric actuator mounted between centers of two legs. The piezoelectric is actuated by a frequency near the natural frequency of the robot. The nonlinear dynamical equations of the system are obtained considering LuGre friction model. To control the system, the equations of motion are linearized using feedback linearization method. Then a stabilizing control law is designed and the robot model is... 

Injuries of neuro-musculoskeletal system impose large costs to the modem societies. Studying the strategies of CNS in planning and control of movements can be helpful in developing diagnoSiS, prognosis and treatment procedures in clinical view and also in designing rehabilitation system and humanoid robots. Tow important features of human movements are redundancy problem and structured variability. Previous studies solved the redundancy problem mainly with optimization methods. Also optimal feedback control theory is the only method that successfully modeled structured variability in voluntary movements. Despite the appropriate outputs this methods are not based on physiological facts... 

The object of this research is modeling, simulation and control of a crystallization process. Population balance equation (PBE) has been used for modeling of such processes. Most of the existing numerical methods cannot solve the general form of PBE in presence of the breakage and agglomeration terms and some others are not suitable for control objectives, because of their intractable CPU time. In this work a new developed numerical method called conservation element and solution element (CE/SE) has been used for solving general form of multidimensional population balance equations. The accuracy and effectiveness of the scheme is demonstrated through computer simulation. According to the... 

Many future space missions, will require a high level of autonomy to service, repair, refuel, reconfigure satellites. As we know there is no project that has demonstrated an actual online pathpelanning for a terminal rendezvous phase in docking maneuver, for an uncooperative tumbling target. Particularly when the target is in a noncircular orbit. Such missions require a 6DOF precise position and attitude tracking maneuver that safely can avoid collision to the target and are capable of flying around. This thesis introduce a new guidance method for vectorial approach to the target in terminal rendezvous phase and applies a Sliding Mode method for attitude tracking control with respect to... 

In this thesis, a bioreactor with oscillatory and chaotic behavior has been studied and a control method has been proposed. This bioreactor shows oscillations for some special values of its parameters. The system dynamic is chaotic when a forced perturbation on the feed concentration is imposed. Dynamic behavior of this oscillatory and chaotic reactor has been illustrated. The main objective is control of product concentration. To do this, first system control is considered by proposing several control strategies for SISO case, then system in MIMO case is studied for achieving a globally controllable system. In SISO case, the performances of controllers are compared based on an objective... 

In this thesis, fuzzy logic is used to suggest a simple algorithm in guidance and control suitable for upper stage decoupler transfer with finite and continuous thrust. It is attempted to improve injection accuracy utilizing classic guidance and control methods and fuzzy logic, while unpredictable errors exist. Orbital perturbations, uncertainties in thrust vector, and initial injection errors in parking orbit, are considered in order to actualize results. In this research, Lambert guidance, velocity to be gained, and fuzzy logics are used for upper stage guidance. Fuzzy-PD control is designed and applied into the upper stage. Final results show acceptable position errors and suitable fuel... 

The satellite attitude control problem is to determine the torque that is produced by actuators in such a way that, the satellite reaches to a desired final state from an initial state. If we do not consider the structural and dynamical equations of actuators and its effects in the system’s equations, the control problem is simple and can be solve easily. But, when the dynamic of the actuators and practical constraints associated with them and the equations of the system are to be considered, due to the dependence of these equations, solving the satellite attitude control is not simple like before. That is why, no effective method is presented to solve it. One of the most widely used... 

The present research intends to design a controller for modular reconfigurable manipulator robots. Manipulators are made up of several links and joints. By attachment of links and joints in different ways, one may make a variety of robot configurations in modular reconfigurable manipulators.
To meet modularity and reconfiguration needs, hybrid control architecture was selected which contains a central supervisory section and distributed controllers located on individual modules. Central supervisory section specifies the trajectory of the robot and other control tasks like position control, speed control and etc. are performed by distributed controllers. In hybrid architecture, dynamic... 

Air-handling unit (AHU) is one of the installations that is responsible to control the temperature and humidity inside a space using the heating, cooling, humidifier and drying air. At present, energy efficiency in building is an important issue to maintain the comfort conditions. One of the most common ways to optimize the performance of industrial installations is the use of control systems, which leads to the reduction in both of the costs and energy consumption. In this research, a multivariable nonlinear dynamic model of the AHU with one zone in the VAV (variable air volume) system for working in summer is considered. At first, introduction, limitations and driving the equations for the... 

Vacuum assisted resin transfer molding (VARTM) is one of the most efficient and low cost process in composite industries. In this process, the preform is covered by a flexible vacuum bag and resin flows into the mold with applying vacuum pressure at the vent. The main objective in VARTM process is preventing voids and dry spots formation in final composite part because these defects may significantly deteriorate the mechanical properties of a structure. Flow front is an important parameter in this process which have a high influence on the void formation. Resin flow control is required to prevent dry spots or micro voids during the VARTM process. In general, the flow control is classified...