| 1. |
- Duraisamy, Palmani, et al.
(författare)
-
Real-time implementation of deep reinforcement learning controller for speed tracking of robotic fish through data-assisted modeling
- 2024
-
Ingår i: Proceedings of the Institution of mechanical engineers. Part C, journal of mechanical engineering science. - : Sage Publications. - 0954-4062 .- 2041-2983. ; 238:2, s. 572-585
-
Tidskriftsartikel (refereegranskat)abstract
- This article proposes real-time speed tracking of two-link surface swimming robotic fish using a deep reinforcement learning (DRL) controller. Hydrodynamic modelling of robotic fish is done by virtue of Newtonian dynamics and Lighthill’s kinematic model. However, this includes external unsteady reactive forces that cannot be modeled accurately due to the distributed nature of hydrodynamic behavior. Therefore, a novel data-assisted dynamic model and control method is proposed for the speed tracking of robotic fish. Initially, the cruise speed motion data are collected through experiments. The water-resistance coefficient is estimated using the least mean square fit, which is then adopted in the model. Subsequently, a closed-loop discrete-time DRL controller trained through a soft actor-critic (SAC) agent is implemented through simulations. SAC overcomes the brittleness problem encountered by other policy gradient approaches by encouraging the policy network for maximum exploration and not assigning a higher probability to any single part of actions. Due to this robustness in the policy learning, the convergence error becomes low in RL-SAC than RL-DDPG controller. The simulation results verify that the DRL-SAC control with data-assisted modelling substantially improves the speed tracking performance. Further, this controller is validated in real-time, and it is observed that the SAC-trained controller tracks the desired speed more accurately than the DDPG controller.
|
|
| 2. |
- Massouh, Bassam, et al.
(författare)
-
Online Hazard Detection in Reconfigurable Plug & Produce Systems
- 2024
-
Ingår i: Flexible Automation and Intelligent Manufacturing: Establishing Bridges for More Sustainable Manufacturing Systems.. - : Springer Nature. - 9783031382406 - 9783031382413 ; , s. 889-897
-
Konferensbidrag (refereegranskat)abstract
- Plug & Produce is a modern automation concept in smart manufacturing for modular, quick, and easy reconfigurable production. The system’s flexibility allows for the configuration of production with abstraction, meaning that the production resources participating in a specific production plan are only known in the online phase. The safety assurance process of such a system is complex and challenging. This work aims to assist the safety assurance when utilizing a highly flexible Plug & Produce concept that accepts instant logical and physical reconfiguration. In this work, we propose a concept for online hazard identification of Plug & Produce systems, the proposed concept, allows for the detection of hazards in the online phase and assists the safety assurance as it provides the hazard list of all possible executable alternatives of the abstract goals automatically. Further, it combines the safety-related information with the control logic allowing for safe planning of operations. The concept was validated with a manufacturing scenario that demonstrates the effectiveness of the proposed concept.
|
|
| 3. |
- Massouh, Bassam
(författare)
-
Planning and Control of Safety-Aware Plug & Produce
- 2024
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- The Plug & Produce manufacturing system is a visionary concept that promises to facilitate the seamless integration and adaptation of manufacturing resources and production processes. The Plug & Produce control system allows for the automatic addition and removal of manufacturing resources, minimizing human intervention. However, the reconfigurability and autonomous decision-making features of Plug & Produce control systems pose challenges to safety design and control functions.In contrast to conventional manufacturing systems with fixed layouts and processes, ensuring safety in Plug & Produce systems is complicated due to the complex risk assessment process, the difficulty of implementing non-restrictive safety measures covering all possible hazards, and the challenge of designing a reliable controller for consistent safe operation.This thesis addresses these challenges through various contributions. It introduces an automatic hazard identification method, considering emergent hazards after reconfiguration. A novel domain ontology is developed, incorporating safety models specific to Plug & Produce systems. The work also proposes a generic, model-based, and automatic risk assessment method, along with a method for the safe execution of plans based on the results of the risk assessment.The results of this research offer benefits to process planners, who are responsible for coordinating the manufacturing processes with product design in the Plug & Produce system. The proposed solution provides tools for process planners to validate their plans and reduces their safety-related responsibilities. The proposed safety assurance method seamlessly integrates into the multi-agent control of Plug & Produce, providing the control system with risk scenarios associated with process plans. This enables proactive and reliable control, effectively avoiding potential risks during system operation.
|
|
