MAINTAINING DEVIATION OF CENTER OF MASS FOR WALKING OF HUMANOID ROBOT WITH KALMAN FILTER AND FUZZY LOGIC CONTROLLER

• Main Author: Aripin
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/36684
• Institution: Institut Teknologi Bandung

Humanoid robots are versatile mechanical devices that have human-like structures. Based on the structures of the humanoid robot can replace several functions and mimic a human behavior such as recognizing objects, taking objects and walking. One of the problems faced in humanoid robots is the ability to walk on sloping surface areas. The walking balance on the robot requires control techniques that can perfect this ability. Balanced walking robots have controlled criteria such as the Center of Mass (CoM) from humanoid robots. In this thesis Bioloid Premium robots that have 18 degrees of freedom (DoF) are used. The balance criteria possessed by Bioloid Premium can be seen from the center of mass to the bobbing and swivel axis so that three criteria are obtained, namely strong stability, weak stability and critical stability. Maintaining the stability of the robot when walking in a sloping area is done by keeping the trajectory from the center of time (CoM) to remain in the stability area. To achieve this goal, the Kalman filter is used to estimate the state of the system and reduce the influence caused by sensor sensitivity, the CoM path error measured using the Inertial Measured Unit (IMU) is presented as input for the Fuzzy Logic Controller control that produces a control signal to correct the angle some robotic joint are dynamic. The experiment results show that the stability control using fuzzy logic was successfully implemented on Bioloid Premium that runs stably with a test scenario in the uphill area of 7.5 degrees and the accuracy of measurements when Bioloid Premium runs in areas without slope that is 94.35% for the pitch axis and 95.16% for roll axis. But the accuracy of measurements decrease when the Bioloid Premium runs on a 2.5 degree inclination area due to disturbance from the effects of gravity so that it has an accuracy value of 83.87%.