Enhancing 3D Trajectory Tracking of Robotic Manipulator Using Sequential Deep Reinforcement Learning with Disturbance Rejection DOI
Saikat Majumder, Soumya Ranjan Sahoo

2022 European Control Conference (ECC), Journal Year: 2024, Volume and Issue: unknown, P. 2512 - 2517

Published: June 25, 2024

Language: Английский

Artificial intelligence-based position control: reinforcement learning approach in spring mass damper systems DOI
Ufuk Demircioğlu, Halit Bakır

Physica Scripta, Journal Year: 2024, Volume and Issue: 99(4), P. 046003 - 046003

Published: Feb. 28, 2024

Abstract This work examines the use of deep Reinforcement Learning (RL) in mass-spring system position control, providing a fresh viewpoint that goes beyond conventional control techniques. Mass-spring systems are widely used many sectors and basic models theory. The novel aspect this approach is thorough examination impact several optimizer algorithms on RL methodology, which reveals optimal tactics. research applies Deep Deterministic Policy Gradient (DDPG) algorithm for continuous action spaces, where actor critic networks important components assessing agent’s performance. agent trained to follow reference trajectory using Simulink environment modeling. study provides insights into learning performance optimization by evaluating training process force-time graphs, reward Episode Manager charts. Furthermore, effect different combinations optimizers examined. outcomes highlight importance selection revealing significant variations times. As result, better understanding relationship between various provided study’s application reinforcement control. results raise possibility more potent methods controlling complex add expanding field at interface theory learning.

Language: Английский

Citations

7

A deep deterministic policy gradient approach for optimizing feeding rates and water quality management in recirculating aquaculture systems DOI Creative Commons
Wael M. Elmessery, Said Elshahat Abdallah, Awad Ali Tayoush Oraiath

et al.

Aquaculture International, Journal Year: 2025, Volume and Issue: 33(4)

Published: March 31, 2025

Language: Английский

Citations

1

Machine Learning, Mechatronics, and Stretch Forming: A History of Innovation in Manufacturing Engineering DOI Creative Commons
Cosmin Constantin Grigoraș, Valentin Zichil, Vlad Andrei Ciubotariu

et al.

Machines, Journal Year: 2024, Volume and Issue: 12(3), P. 180 - 180

Published: March 7, 2024

This review focuses on the complex connections between machine learning, mechatronics, and stretch forming, offering valuable insights that can lay groundwork for future research. It provides an overview of origins fundamentals these fields, emphasizes notable progress, explores influence fields society industry. Also highlighted is progress robotics research particularities in field sheet metal forming its various applications. paper presenting latest technological advancements integrations from their beginnings to present days, providing into directions.

Language: Английский

Citations

4

Optimizing Deep Reinforcement Learning for Adaptive Robotic Arm Control DOI
Jonaid Shianifar, Michael Schukat, Karl Mason

et al.

Communications in computer and information science, Journal Year: 2025, Volume and Issue: unknown, P. 293 - 304

Published: Jan. 1, 2025

Language: Английский

Citations

0

Enhanced toolface angle control of stabilized platform using I_DDPG in rotary steerable system DOI
Aiqing Huo, Kun Zhang, Xue Jiang

et al.

International Journal of Machine Learning and Cybernetics, Journal Year: 2025, Volume and Issue: unknown

Published: March 7, 2025

Language: Английский

Citations

0

Machine Learning for Battery Energy Storage System (BESS) DOI
S. Elango

Advances in computational intelligence and robotics book series, Journal Year: 2025, Volume and Issue: unknown, P. 169 - 202

Published: Feb. 28, 2025

The progress of technology necessitates the development Battery Energy Storage Systems (BESS) to have improved performance, longer life, higher dependability, and more intelligent management strategies. A significant acceleration calculations, capturing complicated mechanisms increase forecast accuracy, optimisation decisions based on full status information are all capabilities that can be achieved with machine learning. This makes it suitable for real-time due computing efficiency possesses. chapter gives an outline later advancements in Machine Learning, focus presentation novel thoughts, strategies, applications learning innovations Systems. also elucidates various aspects challenges, discuss potential solutions, future avenues exploration Learning within

Language: Английский

Citations

0

A New Proposal for Intelligent Continuous Controller of Robotic Finger Prostheses Using Deep Deterministic Policy Gradient Algorithm Through Simulated Assessments DOI Creative Commons
Guilherme de Paula Rúbio, Matheus Carvalho Barbosa Costa, Claysson Bruno Santos Vimieiro

et al.

Robotics, Journal Year: 2025, Volume and Issue: 14(4), P. 49 - 49

Published: April 14, 2025

To improve the adaptability of hand prosthesis, we propose a new smart control for physiological finger prosthesis using advantages deep deterministic policy gradient (DDPG) algorithm. A rigid body model was developed to represent as training environment. The geometric characteristics and physical properties available in literature were assumed, but joint’s stiffness damping neglected. standard DDPG algorithm modified train an artificial neural network (ANN) perform two predetermined trajectories: linear sinusoidal. ANN evaluated through use computational that simulated functionality prosthesis. demonstrated capacity successfully execute both sinusoidal trajectories, exhibiting mean error 3.984±2.899 mm trajectory 3.220±1.419 trajectory. Observing torques, it found used different strategies movement order adapt trajectories. Allowing combination our able trajectories differed from purely sinusoidal, showing its ability finger. results showed possible develop controller multiple which is essential provide more integrated personalized prostheses.

Language: Английский

Citations

0

A Reinforcement-Learning Approach to Control Robotic Manipulator Based on Improved DDPG DOI
Saikat Majumder, Soumya Ranjan Sahoo

Published: Dec. 18, 2023

One of the exciting development in recent decades has been capacity to teach robots using Rein-forcement Learning (RL) techniques execute certain tasks. Deep Deterministic Policy Gradient (DDPG) is one those RL techniques. This paper proposes an adaptive robust controller based on improved DDPG algorithm for position and velocity control n-link robotic manipulator, which nonlinear uncertain. The designed takes care model nonlinearities, uncertainties also time-varying external disturbances. multiple actor networks a critic network. A reward function proposed order guarantee stable effective learning agent. Finally, numerical simulation are performed two-link robot manipulator as example. outcomes demonstrate robustness, adaptability trajectory tracking accuracy this approach. Neural Lyapunov stability shown controller.

Language: Английский

Citations

1

Integrated Robotic Arm Control: Inverse Kinematics, Trajectory Planning, and Performance Evaluation for Automated Welding DOI Creative Commons

Arif Nur Huda,

Dwi Pebrianti,

Zainah Binti MD. Zain

et al.

Asian Journal Science and Engineering, Journal Year: 2023, Volume and Issue: 2(2), P. 82 - 82

Published: Dec. 6, 2023

This research delves into the automated functionality of robotic arm manipulators, a hallmark Industry 4.0, within manufacturing sector. The study focuses on precise movement adhering to predetermined trajectories, addressing vital aspects inverse kinematics and trajectory planning in control. Utilizing Matlab toolbox, conducts simulations kinematic planning. An experimental setup involving controlled by an Arduino Mega 2560 microcontroller, embedded with algorithm planning, is executed. Data acquisition involves inputting coordinates orientation for automatic welding along straight path. Joint angles are measured using rotary encoders converted Cartesian determine end-effector's position. Discrepancy analysis compares joint simulation values, revealing error margins. Movement quality assessed through Capability Processes (CP) evaluation. Results indicate disparities between simulated values. At input (400mm, 0mm, 300mm), angle errors 2.51º, 0.98º, 1.48º observed joints 2, 3, 5, respectively. Similarly, at (300mm, 1.17º, 1.5º, 2.7º registered same joints. Trajectory during reveals average 2.25 mm 10.57 x y axes. Mean absolute 5 1.9º, 0.48º, 1.91º. Keywords: Robotic Arm Manipulators, Inverse Kinematics, Error Analysis

Language: Английский

Citations

1

Enhancing 3D Trajectory Tracking of Robotic Manipulator Using Sequential Deep Reinforcement Learning with Disturbance Rejection DOI
Saikat Majumder, Soumya Ranjan Sahoo

2022 European Control Conference (ECC), Journal Year: 2024, Volume and Issue: unknown, P. 2512 - 2517

Published: June 25, 2024

Language: Английский

Citations

0