Reinforcement Learning-Based Adaptive Control for Collaborative Robots in Dynamic Environments
Summary:
Considering the mass customization generating paradigm of today, neither robots nor human beings can effectively do all assembly jobs on their own. In order to get around it, collaboration manufacturing involving humans and robots has a lot of promise for ensuring that human processes remain flexible while robot help is very reliable. To do the jobs safely and effectively, it might be difficult to develop a happy existence with people and robots. The relevance of human-robot cooperative work has increased as AI and robotics technologies have developed. Through input from a skilled human trainer, robots are able to master target tasks through interaction reinforcement learning methods. But generally interactive reinforcement learning technology necessitates an additional step to include the trainer's input into the educational dataset, which makes immediate automatic acquisition of novel assignments from humans difficult. Additionally, Collaborative Robots the kinds of reinforcement statements that trainers may employ. Collaboration between humans and robots in a shared environment necessitates a great deal of communication.
The research on intentional behavior detection is driven by the need to distinguish between a coping strategy for an incidental encounter and an active cooperation, since they often have opposing effects. Human-robot interaction (HRI), which is designed to encourage productive cooperation between people and robots in a variety of circumstances, is a quickly expanding topic as a result of the increasing prevalence of service robots in daily surroundings. Due to deep reinforcement learning's (DRL) remarkable accomplishment, scholars began focusing on the multi-agent domain. They absorbed the DRL technique into a system of multiple agents with the goal of accomplishing many intricate tasks in a multi-agent setting, giving rise to reinforcement learning. Additionally, studies have been conducted to teach robots how to communicate with people in a variety of real-world settings and acquire new objective activities. In order for people and robots to collaborate on specialized jobs in an ever-evolving setting, a learning system that promotes natural human-robot interaction is necessary, allowing non-specialists to educate robots. Specifically, person-in-the-loop, a recently developed method, incorporates human expertise into the system.
This special issue presents important topics in cooperative construction and gives a thorough account of the environment in which humans and robots cohabit. The assembly task of complicated reinforcement learning methods is represented and as a result, it is difficult for the machine and its operator to communicate and work together effectively in the shared space.
List of Interested topics include, but are not limited to, the following:
- An adaptive human-robot collaborative fabrication technique facilitated by graph-based reinforcement learning.
- A flexible method for collaborative human-robot production enabled by graph-based reinforcement education.
- A versatile approach to cooperative human-robot manufacturing facilitated by reinforcement learning.
- A deep reinforcement learning-based cooperative control approach for dual-arm robots.
- Sensitivity learning-based adaptation for interacting humans and robots
- Routing using deep reinforcement learning in unfamiliar dynamic surroundings.
- Combining reinforcement learning with variable control to enable immediate oversight and learning.
- Collaborative multi-robot navigation using extensive reinforcement learning in an evolving setting.
- Adaptive optimum control and secure reinforcement learning: implications to avoiding obstacles.
- Supervision based on synchronised reinforcement learning for cognitive sovereignty.
- Scalable deep reinforcement learning combined with policy transmission and adaptive robotics.
Special Issue Guest Editors:
Dr. Mahmud Iwan Solihin
Associate Professor at Faculty of Engineering,
UCSI University, Malaysia mahmudis@ucsiuniversity.edu.my
Dr. Lin Guoping
Professor of Department of Engineering,
Department of Industrial Engineering and Enterprise Information,
Tunghai University, Taiwan kplin@thu.edu.tw
Dr. Slamet Riyadi
Associate Professor, Department of Information Technology,
Universitas Muhammadiyah Yogyakarta, Indonesia riyadi@umy.ac.id
Deadlines:
Article Submission Deadline - [10.11.2024]
Authors Notification Date - [20.02.2025]
Revised Papers Due Date - [25.04.2025]
Final notification Date - [20.06.2025]
