A Machine-Learning-Enabled, Exoskeleton that Optimizes Walking Rate and Energy Efficiency

This exoskeleton uses machine learning to put a personalized spring in your step
In a recent press release, Steve Collins, an associate professor of mechanical and electrical engineering at Stanford University who heads the Stanford Biomechatronics Laboratory said, \”This exoskeleton provides assistance to people as they walk through the real-world.\” It resulted to exceptional improvements in walking efficiency and speed.

The team developed a machine-learning algorithm that allows for personalization. They trained it using emulators, or machines that collect data from volunteers hooked up to the machines. The volunteers walked in different speeds to simulate scenarios such as trying to catch a train or strolling through a park.

The algorithm made connections between the scenarios and energy expenditure of people, using these connections to learn how to assist wearers in a manner that is actually helpful to them. The algorithm measures how a person’s movements change when they put on the boot. It tests different patterns of assistance every time. The algorithm has a relatively short learning curve and can adapt to new users within an hour.


This Exoskeleton Uses AI to Help People Walk Faster With Less Energy