The Technology Innovation Institute (TII), part of the Advanced Technology Research Council in Abu Dhabi, has launched RLtools, an open-source reinforcement learning library.
This initiative aims to provide effective solutions to some of the most pressing training challenges in the field of autonomous systems.
The library is a joint project developed by the TII’s Autonomous Robotics Research Center team and the Robotics and Perception Lab at New York University (NYU). This project marks a significant milestone titled “Learning to Fly in Seconds” as it is the first time that high-speed drone controllers have been trained using a standard, commercially available computer.
The collaboration between the Technology Innovation Institute and New York University highlights the exceptional research capabilities of both institutions and underscores their shared commitment to innovation and flexibility.
Both parties aim to create a versatile and adaptable framework that will expand the range of algorithms available in the library, enhance precision, and broaden compatibility across different platforms, whether in simulated environments or real-world applications.
The ultimate goal of the project is to develop a unique and comprehensive control unit capable of autonomous operations and instantaneous learning.
It also aims to dynamically adjust its parametric measures to adapt to surrounding conditions. This approach is designed to establish a unified system capable of addressing the challenges and variable factors presented by diverse environments with maximum accuracy and efficiency.
Dario Albani, Senior Director at the Autonomous Robotics Research Center at TII, commented, “Our close collaboration with New York University has led to the launch of the RLtools open-source library. We are confident that it will drive progress in the fields of reinforcement learning and continuous control by reducing the time required for training and providing a flexible framework. RLtools will catalyse rapid and impactful advancements in autonomous systems intelligence.”
Professor Giuseppe Loianno, Assistant Professor at New York University and Director of the Robotics and Perception Lab, added, “RLtools demonstrates the significant strides we have made in developing the next generation of practical and efficient autonomous systems in terms of resource consumption and adaptability. By continuing our research efforts and close collaboration with the Autonomous Robotics Research Center, RLtools aims to offer more innovative solutions, making reinforcement learning a fundamental aspect of designing intelligent systems that tackle future challenges.”
This joint effort represents a significant advancement in the field of artificial intelligence and autonomous robotics, paving the way for more sophisticated and adaptable autonomous systems in the future.
