Fuzzy-logic based gripper design and analysis for objects grasping

• Main Author: Nor Anis Aneza, Lokman
• Format: Thesis
• Language: English
• Published: 2017
• Subjects: TK Electrical engineering. Electronics Nuclear engineering
• Online Access: http://umpir.ump.edu.my/id/eprint/19777/19/Fuzzy-logic%20based%20gripper%20design%20and%20analysis%20for%20objects%20grasping.pdf; http://umpir.ump.edu.my/id/eprint/19777/
• Institution: Universiti Malaysia Pahang

Three fingered gripper is one of the examples for end effector in robotic technologies that are actuated by the controller and joint actuator. Gripper has been used widely in industrial applications such as to pick and place, product assembly, paintings, and other tasks. In this research, one of the methods in Artificial Intelligence known as Fuzzy Logic (FL) is applied to control the gripper movement for object grasping. Firstly, the gripper is designed by using SimMechanics in Matlab Simulink to get the better outputs torque and simulated angle in grasping technique. The grasping of different sizes of objects are also examined and then the comparison between two controllers; FL and Proportional-Integral-Derivative (PID) were organized to assess their performance. The methods of FL with feedback controller (FLwFB) is proposed in this research to compare its performance to the FL without feedback (FLw/oFB) in the system. Simulation analysis for this comparison is also presented. The gripper grasps the objects properly with the feedback to adjust the value in the system. The system has two inputs; desired angles and sizes of objects and two outputs; simulated angles and torques. The value of desired angles are compared with the simulated angles as a reference for the analysis purposes. This measurement demonstrates the gripper condition after grasping an object. Findings have shown that the outputs for trapezoidal function in FL has better performance than the other membership functions especially when the time response is concerned. During the investigation, there are two types fuzzy rules are taken in the analysis which is 3 number of rules and 5 number of rules. Simulation results suggested that 3 number of rules demonstrated a better solutions for this proposed system. The comparison results between FL and non-FL are also analyzed in the research and outcomes illustrated that FLwFB is better than others; the error between desired and simulated angle is small when the gripper grasping an object. To evaluate the performance of the proposed technique, FL and PID controllers performance in gripper are also considered. The result indicates that FL surpassed the PID performance when time response is considered. As a result of above research works, a new contribution for FL based gripper design has been achieved for different kind of object with a simple feedback system.