Robot Path Planning via Deep Reinforcement Learning with Improved Reward Function

Guo, Meishan; Wang, Yao; Liang, Binyan; Chen, Zhihong; Lin, Junqin; Huang, Kui

doi:10.1007/978-981-16-6320-8_68

Cited by 3 publications

(3 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…One crucial aspect of DRL algorithms is the reward function, which fundamentally shapes the agent's learning strategy and the direction for network optimization. Crafting an ideal explicit reward function to meet long-term goals is a formidable task, chiefly because mapping relationships from complex state spaces to reward values can be nonlinear, making the manual description of the relationships between reward components highly challenging [33]. Initial research focused predominantly on single-objective optimization, primarily centered on position control [34,35], simplifying the reward function as follows:…”

Section: Cascaded Fuzzy Reward System (Cfrs)mentioning

confidence: 99%

Cascaded Fuzzy Reward Mechanisms in Deep Reinforcement Learning for Comprehensive Path Planning in Textile Robotic Systems

Zhao,

Ding,

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

With the rapid advancement of industrial automation and artificial intelligence technologies, particularly in the textile industry, robotic technology is increasingly challenged with intelligent path planning and executing high-precision tasks. This study focuses on the automatic path planning and yarn-spool-assembly tasks of textile robotic arms, proposing an end-to-end planning and control model that integrates deep reinforcement learning. The innovation of this paper lies in the introduction of a cascaded fuzzy reward system, which is integrated into the end-to-end model to enhance learning efficiency and reduce ineffective exploration, thereby accelerating the convergence of the model. A series of experiments conducted in a simulated environment demonstrate the model’s exceptional performance in yarn-spool-assembly tasks. Compared to traditional reinforcement learning methods, our model shows potential advantages in improving task success rates and reducing collision rates. The cascaded fuzzy reward system, a core component of our end-to-end deep reinforcement learning model, offers a novel and more robust solution for the automated path planning of robotic arms. In summary, the method proposed in this study provides a new perspective and potential applications for industrial automation, especially in the operation of robotic arms in complex and uncertain environments.

show abstract

Section: Cascaded Fuzzy Reward System (Cfrs)mentioning

confidence: 99%

Cascaded Fuzzy Reward Mechanisms in Deep Reinforcement Learning for Comprehensive Path Planning in Textile Robotic Systems

Zhao,

Ding,

et al. 2024

Applied Sciences

View full text Add to dashboard Cite

show abstract

“…Examples of the RL-based algorithms utilized in motion planning include the twin delayed deep deterministic policy gradient (TD3) [107] and Exploitation of Abstract Symmetry of Environments (EASE). The latter relies on locally adopting spatial symmetry abstractions obtained from naïvely trained agents [108].…”

Section: Reinforcement Learningmentioning

confidence: 99%

“…As aforementioned, training RL models is challenging in terms of convergence and robustness. The reasons for this include ambiguities in the relationship between the Cartesian and joint spaces, continuous workspaces, and redundant DoF, which result in unnecessary explorations [107]. This could be alleviated using a NN to produce an initial policy for guiding the training of the RL framework [105].…”

Section: Reinforcement Learningmentioning

confidence: 99%

Conventional, Heuristic and Learning-Based Robot Motion Planning: Reviewing Frameworks of Current Practical Significance

2023

View full text Add to dashboard Cite

Motion planning algorithms have seen considerable progress and expansion across various domains of science and technology during the last few decades, where rapid advancements in path planning and trajectory optimization approaches have been made possible by the conspicuous enhancements brought, among others, by sampling-based methods and convex optimization strategies. Although they have been investigated from various perspectives in the existing literature, recent developments aimed at integrating robots into social, healthcare, industrial, and educational contexts have attributed greater importance to additional concepts that would allow them to communicate, cooperate, and collaborate with each other, as well as with human beings, in a meaningful and efficient manner. Therefore, in this survey, in addition to a brief overview of some of the essential aspects of motion planning algorithms, a few vital considerations required for assimilating robots into real-world applications, including certain instances of social, urban, and industrial environments, are introduced, followed by a critical discussion of a set of outstanding issues worthy of further investigation and development in future scientific studies.

show abstract

A review of recent trend in motion planning of industrial robots

Tamizi

Yaghoubi

Najjaran

2023

Int J Intell Robot Appl

View full text Add to dashboard Cite

Robot Path Planning via Deep Reinforcement Learning with Improved Reward Function

Cited by 3 publications

References 8 publications

Cascaded Fuzzy Reward Mechanisms in Deep Reinforcement Learning for Comprehensive Path Planning in Textile Robotic Systems

Cascaded Fuzzy Reward Mechanisms in Deep Reinforcement Learning for Comprehensive Path Planning in Textile Robotic Systems

Conventional, Heuristic and Learning-Based Robot Motion Planning: Reviewing Frameworks of Current Practical Significance

A review of recent trend in motion planning of industrial robots

Contact Info

Product

Resources

About