A roadmap for end-to-end task-agnostic exoskeleton control
Nature Machine Intelligence, Published online: 24 August 2026; doi:10.1038/s42256-026-01297-7 Shepherd et al. propose end-to-end AI control of lower-limb exoskeletons based on real-time estimates of physiological signals.
The recent article in Nature explores the potential of end-to-end, AI-driven control systems for lower-limb exoskeletons. Unlike traditional methods that rely on discrete task classification and struggle to adapt to the continuous, variable nature of human movement, these new systems estimate real-time physiological states, particularly joint moments.
The researchers highlight several key challenges: the need to optimize these systems, incorporate safety mechanisms, and reduce the time required to collect suitable training datasets. They also discuss recent advances in this field, including the development of soft robotic shorts for increased outdoor walking efficiency in older adults and the use of deep convolutional neural networks for environment classification in robotic leg prostheses and exoskeletons.
The upcoming systems promise to meaningfully augment human mobility across various populations, environments, and devices, potentially transforming how we assist individuals with mobility decline, disabilities, and workplace injuries.
Written by urgent.news from Nature Machine Intelligence's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
