Performance Improvement in Robotic Arm Movement using Fuzzy and Genetic Algorithms

Abstract

Robotics is no more restricted to defined steps of machine rather it becomes more independent and self-decision making machine with the thinking power of artificial intelligence. Robots are used in the industrial applications like assembly operations, spray painting, grinding, welding, pick-and-place operations, etc. Robots perform repetitive tasks with accuracy and higher efficiency. Fully autonomous robots are still under research, which can take decision on their own to perform any task and are able to perform tasks efficiently. Industrial robots are quite prevalent in high volume manufacturing processes. In many field applications where technical support is required, manhandling is either dangerous or is not possible. In such situations, three or more arm manipulators are commonly used. They are in great demand to speed up the automation processes. Two-link robotic arm should be able to locate any location, which is the required movement in real world situations. These are used in micro to macro scale applications, viz., chip fabrications to huge mechanical actuators used in chemical processes. In these cases, the motion profile of the robot rarely changes throughout the whole operation. Therefore, searching an optimal robot arm movement is a favorable solution to those problems. Human beings are extremely complicated systems. The researcher’s aim is to develop autonomous robot, which can compete with them on any level, and to perform tasks as human being perform with accuracy and efficiently. Path planning is one of the important objectives of developing autonomous robots. The evidence of the success of the intelligent human path planner suggests that the use of some intelligent, automated motion planning would be useful to investigate. Artificial intelligence can make it possible for robots to perform complex tasks easily. Advance computers have shown themselves to be extremely capable in many application areas, and have transformed the world in which we live. Their application has helped to enhance the quality of traditionally human tasks, and has completely automated other tasks, replacing the need for humans to carry them out. However, automated systems controlled by computers cannot beat men in all areas. The advancement in computational techniques has followed a new branch called as Artificial Intelligence namely evolutionary computation. Techniques falling in the evolutionary computation area are also referred to as evolutionary algorithms (EA). Evolutionary computation has become the standard umbrella for a number of evolutionary driven techniques. This research work is restricted to path planning of robotic arm movement with the help of fuzzy and genetic algorithms in order to achieve the target of minimum energy consumption criterion. Existing techniques require much computation and are dependent upon the mathematical model accuracy. In every relevant field, it is desired to minimise the consumption of power by the appliance. This can be achieved by optimising the movement of robotic arms. This research work has been implemented on 2 degree-of-freedom (DOF), 3 DOF and 4 DOF robotic arm manipulators, which can be generalised to more degree of freedom manipulators. Artificial techniques like fuzzy, genetic algorithms and neural networks are widely used in various fields of sciences including robotics. Here, fuzzy and genetic algorithms have been used in hybrid mode. These two tools have been successfully combined in order to maximize their individual strengths so as to achieve the target of optimisation of arm movements between an initial and a final goal positions. The increase in the publications using artificial techniques for various engineering applications acclaimed superiority over other techniques. In this research work, new method based on fuzzy and genetic algorithms has been developed. It provides a probabilistic approach to the path planning problem. The simulation results for various degree-of-freedom have been demonstrated and also issues related to them are discussed. This research work contributes towards the significant role of fuzzy and genetic algorithms techniques in the future path planning methods so as to develop autonomous robots. With increase in degree-of-freedom and higher dimensional work spaces, complexity increases. It become very difficult to find the solutions of path planning problems with classical or mathematical techniques. It is also very difficult to develop accurate mathematical model. Genetic algorithms and fuzzy-genetic algorithms play vital role in path planning problem. This research work is limited to few degree-of-freedom (DOF), however, it may extended to more degree-of-freedoms (DOFs). In this research work, the non-deterministic parameters like friction, settling time and movement have been compensated. These parameters have been successfully compensated and minimised for their effects on the performance of robotic arm movements in order to achieve the target effectively. There may be more number of non-deterministic parameters that may be compensated with little modifications in the future research work. Genetic algorithms and fuzzy-genetic algorithms have carried out an important role in path planning. This research work contributes to the ultimate goal of the autonomous robots and provides motivation to the other researches in this direction. This thesis is divided into seven chapters. A brief outline of each chapter is given here below. The first chapter introduces the issues related to robotics. These robots have to work in repetitive manner in a highly structured predictable environment. The robots are supplied with inputs like position, orientation and complex work spaces in various industrial environments with required jigs and fixtures. It requires large time and incurs huge costs. Such an industrial application needs to be properly installed and accurately calibrated for the spatial specifications of these jigs and fixtures. Attempt has been made to develop artificially intelligent robotic systems, which will sense their external environment and operate on the basis of known information for which they have been prepared before hand. Developing an automated robotic system is a challenging task. It requires various skills like programming, mechanical modeling of a robotic system, electronic devices, their interfaces and control. An integrated design based on above skills and relevant engineering areas will help in realisation of such an automated robotic systems. Fuzzy logic and genetic algorithms have been discussed. A brief outline of analytical hierarchy process and hybrid algorithm is also discussed. In second chapter, a detailed review of the literature relating fields of fuzzy logic and genetic algorithms followed by its applications in the engineering field and especially for robotics and automation has been given. The third chapter discusses overview of the robotics specifications based on work envelop geometries and degree-of-freedom. The robotic manipulators and their kinematics have been elaborated for direct kinematics and inverse kinematics. The detailed overview of conventional approaches of path planning has also been presented. Chapter four describes detailed overview of arm movements of 2 DOF, 3 DOF and 4 DOF robotic arm manipulators and evaluates the mathematical model of the arm, while considering case study of these manipulators for calculating various values for joint angles using inverse kinematics. Chapter five presents the implementation of fuzzy logic for arm movement and genetic algorithms for obtaining optimal solution from possible solutions for path optimisation problem. In this chapter, detailed steps involved in implementation of fuzzy logic and genetic algorithms have been given. Simulation results of 2 DOF, 3 DOF and 4 DOF robotic arm manipulator are presented so as to justify the proposed technique, which is being used for path optimisation, while achieving the target of minimum energy consumption. In chapter six, results are shown for 2 DOF, 3 DOF and 4 DOF robotic arm manipulator with genetic algorithms and fuzzy–genetic algorithms. In this chapter, comparison of results with genetic algorithms and fuzzy–genetic algorithms have been discussed. Chapter seven presents the conclusions drawn from this research. It discusses the advantages of the techniques used in this work. Also, some suggestions are made for further research in this area.