Identification and Characteristics of Parallel Actuation Robot’s Leg Configuration Using Hammerstein-Wiener Approach

• Main Authors: Guni, Geogia, Addie Irawan, Hashim
• Format: Article
• Language: English
• Published: Penerbit UMP 2016
• Subjects: Q Science (General); TJ Mechanical engineering and machinery
• Online Access: http://umpir.ump.edu.my/id/eprint/11713/1/JEECIEV1N10.pdf; http://umpir.ump.edu.my/id/eprint/11713/; http://apps-cfm.ump.edu.my/research/jeecie/manuscript/267/JEECIEV1N10.pdf
• Institution: Universiti Malaysia Pahang

This paper presents the modeling of a leg of Quadruped with Parallel Actuation Leg (QPAL) robot. QPAL leg designed with 3 Degree of Freedom (DOF) configuration with indirect or parallel actuation for each joint mimicking a muscle of life form creature such as insect and bugs classified into shoulder, thigh and shank parts. Indirect actuation configuration on its leg makes this robot has different perspective on joint rotational drive and control. Therefore, this project has taken initiative to identify and modeling this indirect actuation joint by using system identification (SI) in order to obtain a mathematical model of each joint of QPAL robot’s leg. A system identification approach was implemented by employing a Hammerstein-Wiener (HW) model as model structure. The state-space model and the transfer function are designed and generated using Hammerstein-wiener modeling procedures start with experiment setup and data collection from experiment. Continue with data processing, selecting model structure, estimation and validation of the model using system SI toolbox in MATLAB®. The best percentage fits for Joint 1, Joint 2 and Joint 3 are 71.06%, 79.14% and 71.35% respectively, meaning that the estimated model is almost tracking the real output data from the experiments. The model for Joint 1 is ideally acceptable and highly applicable since the correlation curves lie between the confidence interval. While the model for Joint 2 and Joint 3 are considered well and acceptable as the correlation curves are almost lies between the confidence interval. The balances 28.94%, 20.86%, and 28.65% are losses due to nonlinear factor such as friction, backlash, torque, and external disturbance.