Fuzzy Dynamic Control of Robotic Arm Manipulator with varying degree of freedom.
Research Field- Major- Electrical Engineering
Minor- Robotics and Control
The purpose of the research project is to develop an effective Fuzzy Logic based controller for a Robotic Arm Manipulator. In industries mostly conventional control techniques are used for controlling a robotic system. Due to high non-linearity and complexity of a robotic system, the performance of the conventional controllers is not very satisfactory. Fuzzy Logic based control has shown better performance in controlling complex and non-linear system in past. For efficient control of robotic arm manipulator, an appropriate dynamic and kinematic model will be developed for moving robotic arm manipulator from initial position to final angular position. For rapid computation and effective control, algorithms will be developed for the selection of inverse kinematic solution and learning of the parameters of the controller. The designed Fuzzy Logic controller could be employed to efficiently control a robotic arm manipulator used in industry to increase its performance or can be used in the robotic system to work in a bio-hazardous and radioactive environment and in turns can save human life working on site.
1. Proposed a simple solution for inverse kinematic problem
2. Effective and efficient dynamic fuzzy controller will be designed.
3. Efficient selection of the inverse kinematic solution.
4. It can be used in bio-hazard and the radioactive environment in turns can save human life working on site.
1. To develop a kinematic and dynamic model of a robotic arm.
2. To translate physical 3-D coordinate into arm angular coordinate.
3. To develop an algorithm for selection for an efficient inverse kinematic solution.
4. To develop an algorithm for parameter learning of dynamic controller.
5. To design a Fuzzy Logic based controller for Robotic Arm Manipulator.
1. The literature review will be conducted.
2. Kinematic modelling of Robotic Arm Manipulator will be carried out.
3. After successful completion of kinematic modelling of Robotic Arm Manipulator, an algorithm for an inverse kinematic solution will be developed.
4. Out of several inverse kinematic solutions, an algorithm will be proposed for the selection of the best inverse kinematic solution.
5. A dynamic model of Robotic Arm Manipulator will be developed for moving the Robotic Arm Manipulator from initial position to final angular position.
6. Dynamic controllers’ parameter learning algorithm will be proposed.
7. Finally, the Fuzzy Logic based controller will be designed to control Robotic Arm Manipulator with varying degree of freedom.
Research Problem and Significance-
In recent times, application of robots in the industries has increased by many folds. The productivity of the industries which utilizes the services of the robots has climbed up considerably. The most common robots employed in industries are used to perform picking and placing objects. Due to the high non-linearity in the robotic system, the performance of the controllers based on conventional techniques is not very satisfactory. Designing an effective controller for industrial robotic arm manipulators can significantly improve the productivity of the industry in which they are profoundly used, which in turns advantageous for a semi-developed country like Saudi Arabia. An extensive research work in the areas of mathematics, control and robotics is needed for designing a Fuzzy Logic based controller for a robotic arm manipulator. There are several well-known techniques present for solving inverse kinematic problems related to the robotics. During the project, a simple and efficient solution for the inverse kinematic problem will be proposed. The method could be utilized for designing many other types of controllers for dynamic systems. The designed controller can be employed in robotic arm manipulators, which can effectively be used in a bio-hazardous and radioactive environment, in turns can save human life working on site.
To design a Fuzzy Logic based controller for Robotic Arm Manipulator (RAM), the very first challenge is to formulate the mathematical equations for the kinematics of a RAM. The kinematics equations of a robot can be divided into two parts, forward and inverse kinematics. These kinematics equations resolve the parameters of the joints of the robotic arm manipulator. Quaternion space and Cartesian space are generally used spaces in kinematics modelling of robotic arm manipulators. There are various methods to transform Cartesian space to Quaternion space, such as Gibbs vector, Euler angles, Pauli Spin matrices etc. Also, Denavit-Hartenberg method is popularly used for defining kinematics of a robot. Once forward kinematics and translation of physical 3-D coordinates into arm angular motion is done, solution for inverse kinematic equations is required. Analytical and Numerical methods are mainly two techniques for finding the solutions for the inverse kinematic problems. In the analytical method, the solution for the joint variables is obtained according to given configuration data. The solution is computationally extensive it generally requires a substantial amount of time to control manipulators in real-time. It would be challenging to develop a novel algorithm for efficient selection of the inverse kinematic solution using numerical methods. Different methods of evolutionary computational techniques could be utilized to develop the algorithm for obtaining solutions. Learning is required for the controller to tune the parameter of the fuzzy controller. Performance of the Fuzzy Logic based controller depends on the efficient learning of its parameters. An algorithm will be developed to guide the evolution process of the fitness function based on the different parameter of the robotic arm manipulator. After obtaining all the parameters and on completion the designing, the Fuzzy Logic based controller will be tested on a robotic arm manipulator.
A robotic arm manipulator is a programmable mechanical arm which can perform numerous tasks such as placing and picking objects like a human arm. The manipulators were basically designed to perform tasks in radioactive environments or for use in inaccessible places. The links of a robotic arm manipulator are connected by joints that allow linear displacement and rotational motion within a given number of degrees of freedom. The most important thing in the advanced control of robotic arm manipulators is to track the trajectories for performing any task. The performance of a RAM depends on its speed, payload weight and precision. Controlling the robotic system with high accuracy and speed is very difficult by conventional techniques because of complexity and high non-linearity of the system. In designing a conventional controller, a linearized model of a process or a control system is taken. Thus, the results yields by these types of controller are not very satisfactory for highly complex and non-linear systems. Also, the reason for the unsatisfactory performance of the conventional controller is that the linearization of the complex non-linear system is often not accurate. On the other hand, the Fuzzy controller has shown great results than those obtained from the conventional controllers in many applications. The fuzzy controller can yield desired results for a non-linear system such as Robotic Arm Manipulator by adjusting some of the controller parameters.
Accurate formulation of mathematical equations is necessary to describe the proper dynamic behaviour of a robotic arm manipulator. Quaternion space and Cartesian space are generally used spaces in kinematics modelling of robotic arm manipulators. Gibbs vector, Euler angles, Pauli Spin matrices etc. are the methods can be used to transform Cartesian space to Quaternion space. Denavit-Hartenberg also has become a standard for defining kinematic equations in robotics, it requires four parameters for general transformation between two joints. Although dual quaternions can present rotation and transformation in a simple vector form, that would be advantageous for computational storage. Other methods for writing equations of motion describing the dynamics of the links and joints of a robotic arm manipulator can be obtained by using Lagrange-Euler(L-E) and Newton-Euler(N-E) laws. These equations can be used to evaluate the performance of the robotic arm manipulator by computer simulation of the control techniques with the kinematic design of the robotic arm manipulator. Formulation of these equations for the dynamic model of a robotic arm manipulator can be obtained for different aims. Rapid computation, facilitate control analysis and synthesis and improve computer simulation of robotic arm motion could be the purposes. An effective control strategy can be developed using L-E equations for the dynamics of robotic arm manipulators. But, a fair amount of time is required for arithmetic operations to compute the dynamic coefficients. Thus, it is quite unlikely to use L-E equations for real-time control purposes unless they are simplified. Also, the d’Alembert principle can be utilized to obtain a set of equations describing the dynamic motion of a robotic arm. It is also easy to develop suitable control strategy as the equations are in vector-matrix form. Developing equations of motions using different methods yields a different order of differential equations. The lower the order of differential equations, the better and efficient control techniques could be developed.
For the Research Duration: (6 Months)
Coordination and distribution of work related to project execution between PI and Co-PI and Research Assistant.
Literature review in detail of the proposed work.
Developing methodology for execution of work.
The dynamic and Kinematic model will be developed using different techniques.
The investigation will be carried out to select a proper model.
The algorithm will be developed for obtaining the solutions for the inverse kinematic problem.
The algorithm will be developed for selecting the best inverse solution from several generated solutions.
· A dynamic model of Robotic Arm Manipulator will be developed for moving the Robotic Arm Manipulator from initial position to final angular position.
· Dynamic controllers’ parameter learning algorithm will be proposed.
· Fuzzy Logic based controller will be designed and tested to control Robotic Arm Manipulator with varying degree of freedom.
The compiling of results and conclusions for writing articles for publication in scientific journals.
Written scientific articles will be transferred for publication in the journals of international repute.
The detailed article will be written for submission to the University’s deanship of research.