Running into the Future: AI-Powered Robotic Exoskeleton Training Breakthrough

Researchers at NC State have developed an AI method to train robotic exoskeletons, enhancing energy efficiency in walking, running, and climbing, promising quicker, personalized robotic assistance.

Researchers at North Carolina State University have unveiled a novel AI-driven method that significantly enhances the functionality of robotic exoskeletons, making them immediately useful in research settings. Detailed in an upcoming issue of Nature, this breakthrough leverages computer simulations to train robotic exoskeletons to autonomously optimize energy efficiency for activities like walking, running, and climbing stairs.

Hao Su, the study’s corresponding author and an associate professor of mechanical and aerospace engineering at NC State, emphasized the importance of this advancement. “This work proposes and demonstrates a new machine-learning framework that bridges the gap between simulation and reality to autonomously control wearable robots to improve mobility and health of humans,” he stated.

The core innovation lies in the method’s ability to bypass traditional, time-consuming human testing phases. By using simulations, the AI embedded within the robotic exoskeleton learns how much force to apply and when, streamlining the process of adapting the technology for practical use. “The key idea here is that the embodied AI in a portable exoskeleton is learning how to help people walk, run or climb in a computer simulation, without requiring any experiments,” Su explained.

Shuzhen Luo, the first author of the paper and former postdoctoral researcher at NC State, now an assistant professor at Embry-Riddle Aeronautical University, highlighted the practical benefits of the new method. “We have developed a way to train and control wearable robots to directly benefit humans,” Luo commented.

The effectiveness of this new approach was underscored by significant energy savings during testing. Participants equipped with the AI-enhanced exoskeletons used 24.3% less metabolic energy while walking, 13.1% less while running, and 15.4% less while climbing stairs compared to performing these tasks unaided.

The implications of this research extend beyond immediate energy efficiency. Su elaborated on the broader potential applications, stating, “Our framework may offer a generalizable and scalable strategy for the rapid development and widespread adoption of a variety of assistive robots for both able-bodied and mobility-impaired individuals.”

The team is also exploring future applications of this technology, particularly for older adults and individuals with neurological conditions like cerebral palsy. “We are in the early stages of testing the new method’s performance in robotic exoskeletons being used by older adults and people with neurological conditions, such as cerebral palsy. And we are also interested in exploring how the method could improve the performance of robotic prosthetic devices for amputee populations,” Su noted..

Looking further afield by allowing robotic exoskeletons to be trained swiftly and efficiently through computer simulations it may be possible to accelerate the development cycle of scientific robotics developed to augment researchers. This would be particularly advantageous in research environments where customizing assistive devices to individual needs or specific research parameters is crucial. Furthermore, the ability to rapidly prototype and test various configurations of assistive robots can lead to more innovative solutions and applications.

This technology not only streamlines the development process but also enhances the potential for personalized healthcare solutions, making it a valuable tool for advancing research in robotic assistance across diverse scenarios.

You can read the full paper “Experiment-free Exoskeleton Assistance Via Learning in Simulation” in Nature

Staff Writer

Our in-house science writing team has prepared this content specifically for Lab Horizons

Leave a Reply

Your email address will not be published. Required fields are marked *