PRESSURE AND BENDING ANGLE CONTROL OF SOFT ROBOTIC USING PID CONTROLLER AND MPC

• Main Author: Fadlankahlil Alifdhyatra, Athar
• Format: Theses
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/68532
• Institution: Institut Teknologi Bandung

Soft robotics use soft materials in its structure which is different from conventional robotics which uses hard materials connected with joints. There are various kinds of soft robotics, one of which is a soft robot with fluid-based actuation that can bend. Due to the various forms and types of actuation of soft robotics, until now there is no common modeling and control method that can be used for soft robotics. In addition, modeling is quite difficult due to the non-linear dynamics of soft robotics and soft materials that have infintite degree of freedom. One of the modeling methods that can be used is the system identification method using experimental data. In this research, a soft robotics control system with pneumatic- based actuation was designed. The system was designed using a linear control system approach in a limited operating area by utilizing a model derived from system identification. The system consists of two closed loops that utilize internal system pressure and soft body bending angle feedback from pressure and bending sensor and actuation of a pump circuit controlled by a microcontroller. The control strategy is applied to the internal air pressure in the system as well as the bending angle of the soft robotic body. The types of controllers used to control the system are PID controllers and MPC. The two controllers are used to provide the pressure reference required by the pressure controller. Experimental results show the ability of the system to track reference input but there are system limitations in maintaining pressure causing the system pressure to oscillate. Based on the test results, a maximum bending angle error of 15 degrees and maximum response time of 30 seconds is obtained for the control system implementations, with implementation of the PID controller providing a better response time but the MPC implementation providing a better bending angle error value.