Guest Editorial: Special Issue on Embodied Intelligence for Wearable Robotics

Wearable robotics have attracted considerable attention due to their potential advantages in human augmentation, intelligent assistance, and enhanced user experiences. Nevertheless, the real-world deployment of wearable robotic systems remains constrained by limitations in intelligence and adaptability, especially under the inter- and intra-user variability encountered in rehabilitation and daily-life scenarios. These challenges can be addressed by integrating advanced artificial intelligence—notably embodied intelligence—to enable wearable robots to perceive, learn, decide, and act in a tightly coupled manner with the human body and environment. This Special Issue on Embodied Intelligence for Wearable Robotics brings together recent advances that span neuromorphic learning paradigms, multimodal intention understanding, ergonomics-aware interaction, and continual adaptation. Collectively, the accepted contributions highlight a core message: progress in wearable robotics increasingly depends on the codesign of sensing and perception, learning and cognition, and control and embodiment, with a strong emphasis on human-centered interaction. A key enabler for embodied intelligence in wearable systems is efficient, temporally precise computation close to the sensorimotor loop. In this context, the article [A1] surveys the emerging landscape of spiking neural networks (SNNs) for rehabilitative wearable robots. By organizing prior work from the perspectives of motion decoding, robot control, and embedded/neuromorphic implementation, the review clarifies both the promise of event-driven intelligence (e.g., low power and real-time responsiveness) and the practical gaps that must be bridged for robust deployment in human–robot interaction. Another central theme is the use of embodied perception for more natural and volitional assistance. The article [A2] advances this direction by leveraging gaze cues to guide locomotion intention recognition for knee–ankle prosthetic control. By fusing eye-gaze information with other sensor modalities, the approach aims to infer user intent earlier and more reliably, thereby improving responsiveness and user experience in complex locomotion transitions. Beyond intention understanding, wearable robotics must also account for comfort, safety, and ergonomics in close physical interaction. The article [A3] addresses this need by investigating ergonomic optimization in robot–human handover scenarios. Although object handover is often studied in collaborative robotics, this work underscores its relevance to wearable and assistive systems where physical interaction quality, user workload, and comfort can directly influence acceptance and long-term usability. Finally, long-term adaptation is indispensable for wearable robots operating across diverse users and changing conditions. The article [A4] presents an adaptive learning framework for hip joint angle prediction that incorporates continual learning. By emphasizing continual adaptation, this work contributes toward personalized wearable assistance that can remain effective despite drift in user gait patterns, device dynamics, or task context. Taken together, the articles in this Special Issue demonstrate how embodied intelligence can elevate wearable robotics from “preprogrammed assistance” toward “context-aware, user-adaptive companionship” in rehabilitation and augmentation. Looking forward, several research opportunities appear particularly promising: 1) neuromorphic and edge intelligence for low-latency, low-power wearable computation; 2) multimodal perception and intent inference that integrate gaze, vision, Electromyography (EMG), kinematics, and interaction forces; 3) ergonomics- and comfort-aware learning objectives that explicitly optimize human well-being; and 4) continual, personalized learning with safety guarantees for long-term human–robot coadaptation. We sincerely thank all authors for their outstanding contributions and all reviewers for their timely, rigorous, and constructive feedback. We also express our gratitude to the Editor-in-Chief and the editorial staff of IEEE Transactions on Cognitive and Developmental Systems for their support throughout the preparation of this Special Issue. We hope this collection will serve as a valuable reference and inspire further advances in embodied intelligence for wearable robotics.