No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Hill Climbing and Q-Learning on a Ball Balancing System
ResearchGate has not been able to resolve any citations for this publication.
Reinforcement learning (RL) has distinguished itself as a prominent learning method to augment the efficacy of autonomous systems. Recent advances in deep learning studies have complemented existing RL methods and led to a crucial breakthrough in the effort of applying RL to automation and robotics. Artificial agents based on deep RL can take selective and intelligent actions comparable to those of a human to maximize the feedback reward from the interactive environment. In this paper, we survey recent developments in the literature regarding deep RL methods for building human-level agents. As a result, prominent studies that involve modeling every aspect of a humanlevel agent will be examined. We also provide an overview of constructing a framework for prospective autonomous systems. Moreover, various toolkits and frameworks are suggested to facilitate the development of deep RL methods. Finally, we open a discussion that potentially raises a range of future research directions in deep RL.
Web-based labs are key tools for distance education that help to illustrate scientific phenomena which require costly or difficult-to-assemble equipment. Easy Java Simulations (EJS) is an authoring tool that speeds up the creation of that kind of labs. An excellent proof of the EJS potential is the Open Source Physics (OSP) repository, which hosts hundreds of free EJS labs. Learning management systems, such as Moodle, provide social contexts where students interact with each other. The work described in this paper looks for the synergy of both tools, EJS and Moodle, by supporting the deployment of EJS labs into Moodle and thus enriching them with social features (e.g., chat, forums, videoconference, etc.). To test this approach, the authors have created the ball and beam lab, which helps students of automatic control engineering to train different advanced techniques (robust, fuzzy and reset control), and compare their performance in relation to a conventional proportional-integral-derivative control.
This work presents an evolutionary controller design method for a ball-and-beam system. The method consists of a population of feedforward neural network controllers that evolve towards an optimal controller through the use of a genetic algorithm. The optimal controller is then applied to several different initial positions from which it has to balance the system. From the simulation results, we can conclude that the evolved neural controller can balance the system effectively.
Deep reinforcement learning is poised to revolutionise the field of AI and represents a step towards building autonomous systems with a higher level understanding of the visual world. Currently, deep learning is enabling reinforcement learning to scale to problems that were previously intractable, such as learning to play video games directly from pixels. Deep reinforcement learning algorithms are also applied to robotics, allowing control policies for robots to be learned directly from camera inputs in the real world. In this survey, we begin with an introduction to the general field of reinforcement learning, then progress to the main streams of value-based and policy-based methods. Our survey will cover central algorithms in deep reinforcement learning, including the deep Q-network, trust region policy optimisation, and asynchronous advantage actor-critic. In parallel, we highlight the unique advantages of deep neural networks, focusing on visual understanding via reinforcement learning. To conclude, we describe several current areas of research within the field.