An enhanced Active Reinforcement Learning for Autonomous Robotics in Industrial automation
- Resource Type
- Conference
- Authors
- Rajan, V. Aravinda; Marimuthu, T.; Bhardwaj, Rajat; Shukla, Rishi Prakash
- Source
- 2023 IEEE 2nd International Conference on Industrial Electronics: Developments & Applications (ICIDeA) Industrial Electronics: Developments & Applications (ICIDeA), 2023 IEEE 2nd International Conference on. :193-198 Sep, 2023
- Subject
- Components, Circuits, Devices and Systems
Fields, Waves and Electromagnetics
Power, Energy and Industry Applications
Robotics and Control Systems
Signal Processing and Analysis
Transportation
Automation
Service robots
Scalability
Collaboration
Reinforcement learning
Manuals
Programming
autonomous
environments
industrial
reinforcement
collaborative
- Language
An enhanced active reinforcement learning technique has been proposed to enable autonomous robots to operate and execute tasks in industrial automation. This approach combine hierarchical reinforcement learning and Bayesian optimization, to acquire knowledge from complex real-world environments and acquire optimal policies which can enable autonomous robots to perform collaborative tasks efficiently. The main advantage of this enhanced active reinforcement learning approach is the capability of the autonomous robot to autonomously adapt its movements and decision-making strategies when new tasks are required. It allows for the robot to explore its environment and learn how to complete tasks optimally while reducing the burden of manual intervention. Moreover, the proposed approach can generalize its knowledge to establish rewarding collaborative behaviors between robots and humans, thus allowing for collaborative human-robot interactions. This will be beneficial in performing industrial automation with robot cooperative tasks and optimize the efficiency of the industrial automation system..