BALANCE CONTROL FOR BIPEDAL ROBOT IN STANDING AND WALKING USING FUZZY LOGIC

• Main Author: Sobirin, Muhammad
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/58025
• Institution: Institut Teknologi Bandung

This research describes a balance control method for the stable standing and walking of a bipedal robot based on Inertia Measurement Unit (IMU) sensor feedback with a kinematic approach. The IMU sensor was used to measure body’s tilt posture of the robot when standing and walking. In this paper, indication of bipedal walking stability was determined based on the tilt posture of robot body. There are several parameters related to the tilt posture of the robot body, namely the pitch angle and roll angle, where the pitch angle represents the front and back tilt movements, while the roll angle represents the right and left side tilt movements. Complementary filter algorithms are added to obtain accurate pitch and roll angle measurements. Fuzzy Logic Controller (FLC) is further designed to evaluate the tilt posture of the robot body to generate appropriate offset angles to be applied one the corresponding joints of the robot. Evaluation of the robot's body tilt posture was carried out during the Double Support Phase (DSP) with joint angles adjustment strategy and ankle positions. The numerical inverse kinematics method was further developed to convert the ankle position offset to the joint angle offset before being applied to the corresponding joint. The performances of the proposed methods were verified by a standing and walking experiments on 18-DOFs biped robot, ElPistolero (Telkom University). From the experimental results, it can be concluded that the proposed FLC is capable of maintaining the balance of the robot in standing and walking conditions, in the range of pitch error values of –15.80° to 14.31° and roll error values of -13.01° to 13.84°. When performing walking movements, the robot can indeed maintain its body position so that it does not fall, but sometimes the soles slip (slip) because because when the roll error is too large, the servo to roll motion (servo AX-12) will not be strong enough to support the robot's body when the legs are swung while Single Support Phase (SSP), as a result the robot can not walk straight.