International Research Journal of Engineering and Technology (IRJET)
e-ISSN: 2395-0056
Volume: 10 Issue: 07 | Jul 2023
p-ISSN: 2395-0072
www.irjet.net
MODEL PREDICTIVE CONTROL BASED JUMPINGOF ROBOTIC LEG ON A PARTICULAR HEIGHT USING REINFORCEMENT LEARNING Amrita Rai-1, Shyam Dwivedi-2 1,2Department of Computer Science and Engineering, Rameshwaram Institute of Technology andManagement,
Dr A PJ Abdul Kalam Technical University, Lucknow, India -------------------------------------------------------------------------***-----------------------------------------------------------------------Abstract — A robotic limb that is controlled by an MPC and trained via reinforcement learning in a real-worldsetting will attempt to learn how to jump beyond a certain height. A dynamic controller, model predictive control regulates each process by resolving constraints. Moreover, reinforcement learning will assist in developing the robotic leg’s capacity to adapt in situations that change rapidly. For regulating the dynamic jumping of a robotic limb through the control design, the authors of this study provide the reinforcement learning with model predictive control approach. In this situation, MPC are primarily utilized to regulate the torque, angles, positions, frictions, force, and other dynamic characteristics on joints. According to this work, hopping might be taught to a robotic leg using model predictive reinforcement learning in a practical environment. Model-based reinforcement learning includes a component called model, which refers to predetermined Kinematics and Dynamics parameters. PPO (proximal policy optimizer) and A2C (advanced actorcritic) algorithms are used in reinforcement learning. Recent years have seen the discovery of predefined forward and inverse kinematics and dynamics parameters that can assist Robotic leg to precisely jump over certain height. The first section is the introduction, which includes a discussion of recent study findings as well as a review of all relevant literature. A brief explanation of each phrase is covered in the second section, while the third section discusses its applications, workings, and formula. The fourth portion is the conclusion, and the final part is a list of sources used in this study.
1. INTRODUCTION 1.1 Robots A Robots are machines that can be programmed by computers and are capable of per- forming a complex series of tasks on their own. An external control device or an internal control system can both control a robot. Even though some of them are made to resemble people, most robots are task-performing machines that put basic usefulness before emotive aesthetics. The design, maintenance, implementation, and use of robots as well as the computer systems that control, sense, and process data for them are all covered by the field of technology knownas robotics. These technologies deal with automated systems that can replace people in hazardous situations or during production processes or that are similar to people in terms of behaviour, cognition, or appearance. Robots’ ability to perform simple tasks in a manner similar to that of humans makes their work simple, quick, and easy. There are far too many domestic robots in our daily lives. People are rapidly growing more dependent on robots because they do household tasks. Some of the most well-liked robots are vacuum cleaners and kitchen robots, but we also have robots that can mow the lawn, clean the windows, and iron your clothes. Robots will someday take over certain human tasks due to their ever-improving skills, but not all. Robotics can presently only automate 25 percent of occupations in unstable, human-dependent industries like nursing and construction. Robots, however, are and always will be dependent on human programming. Robots are automated machines that require little or no human assistance to complete specified jobs quickly and accurately. Over the past 50 years, there have been significant advancements achieved in the field of robotics, which is focused on the design, creation, and application of robots. Essentially, there are as many different types of robots as there are tasks for them to perform. While some tasks are better performed by people rather than machines, others are better left to them. Robots are more adept at the following tasks than humans are: 1.
In business or industrial settings, automate manual or repetitive processes.
© 2023, IRJET
|
Impact Factor value: 8.226
|
ISO 9001:2008 Certified Journal
|
Page 71