Self-Evolving Cognitive Environments for Autonomous Robotics: A Closed-Loop Learning Architecture for Adaptive Physical Systems
Keywords:
Autonomous Robotics, Closed-Loop Learning Architectures, Policy Evolution, Environmental Adaptation, and System-Level Co-Adaptation.Abstract
Today's autonomous robotic systems are designed to work under heavy architecture constraints․ Although they are well-suited for static assumptions about the environment, most use cases, be it in warehouses, transportation systems, manufacturing, or shared spaces, require robots to be able to adapt to fixed or slowly changing environments. Self-Evolving Cognitive Environments (SECE), on the other hand, relies on system-level architectural advances, wherein the robotic agents, the physical environments, and the artificial intelligence systems function as a single closed-loop system wherein learning is not confined to the agent only but can extend to co-adaptation of the environment dynamics and the agent's behaviors through perception, behavioral modeling, and policy experimentation. This involves a six-layer architecture wherein multi-modal perception, environmental state representation, behavior modeling, closed-loop experimentation, policy learning, and edge-cloud orchestration are tightly interlinked and allow the environment to adapt and improve the operational policy online in the physical environment, rather than relying on offline training in simulated environments. A range of applications (including those in warehouse logistics, autonomous driving, delivery robots, and industrial robotics) shows how system-level learning can lead to strong, scalable, contextual autonomy through the co-adaptation of the environment and the agent as a foundational building block towards capabilities that were previously out of reach for autonomous agents.
DOI: https://doi.org/10.17762/ijisae.v14i1s.8233
Downloads
References
Richard S. Sutton and Andrew G. Barto, "Reinforcement Learning: An Introduction," MIT Press, 2015 [Online]. Available: https://web.stanford.edu/class/psych209/Readings/SuttonBartoIPRLBook2ndEd.pdf
Sebastian THRUN et al., "Probabilistic Robotics," MIT Press, 2000. [Online]. Available: https://docs.ufpr.br/~danielsantos/ProbabilisticRobotics.pdf
David Ha, Jürgen Schmidhuber, "World Models," Arxiv, 2018. [Online]. Available: https://arxiv.org/abs/1803.10122
Mariusz Bojarski, "End-to-End Learning for Self-Driving Cars," arXiv, 2016. [Online]. Available: https://arxiv.org/abs/1604.07316
Chelsea Finn, "Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks," ACM Digital Library, 2017. [Online]. Available: https://dl.acm.org/doi/10.5555/3305381.3305498
Timothy P. Lillicrap et al., "Continuous control with deep reinforcement learning," Arxiv, 2016. [Online]. Available: https://arxiv.org/abs/1509.02971
Max Jaderberg et al., "Reinforcement learning with unsupervised auxiliary tasks," Arxiv, 2016. [Online]. Available: https://arxiv.org/abs/1611.05397
Jürgen Schmidhuber, "Deep Learning in Neural Networks: An Overview," ScienceDirect, 2015. [Online]. Available: https://www.sciencedirect.com/science/article/abs/pii/S0893608014002135
Volodymyr Mnih et al., "Asynchronous methods for deep reinforcement learning," arxiv, 2016. [Online]. Available: https://arxiv.org/abs/1602.01783
John Schulman et al., "Trust region policy optimization," arxiv, 2017. [Online]. Available: https://arxiv.org/abs/1502.05477
Alex Krizhevsky et al., "ImageNet classification with deep convolutional neural networks," ACM Digital Library, 2017. [Online]. Available: https://dl.acm.org/doi/10.1145/3065386
Joseph Redmon, Ali Farhadi, "YOLO9000: Better, Faster, Stronger," arxiv, 2016. [Online]. Available: https://arxiv.org/abs/1612.08242
Marius Cordts et al., "The Cityscapes dataset for semantic urban scene understanding," IEEE Xplore, 2016. [Online]. Available: https://ieeexplore.ieee.org/document/7780719
Radhakrishna Achanta et al., "SLIC superpixels compared to state-of-the-art superpixel methods," PubMed, 2012. [Online]. Available: https://pubmed.ncbi.nlm.nih.gov/22641706/
Kaiming He et al., "Deep Residual Learning for Image Recognition," arxiv, 2015. [Online]. Available: https://arxiv.org/abs/1512.03385
Downloads
Published
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
All papers should be submitted electronically. All submitted manuscripts must be original work that is not under submission at another journal or under consideration for publication in another form, such as a monograph or chapter of a book. Authors of submitted papers are obligated not to submit their paper for publication elsewhere until an editorial decision is rendered on their submission. Further, authors of accepted papers are prohibited from publishing the results in other publications that appear before the paper is published in the Journal unless they receive approval for doing so from the Editor-In-Chief.
IJISAE open access articles are licensed under a Creative Commons Attribution-ShareAlike 4.0 International License. This license lets the audience to give appropriate credit, provide a link to the license, and indicate if changes were made and if they remix, transform, or build upon the material, they must distribute contributions under the same license as the original.
