CFP-WMP: A Coarse-to-Fine Environmental Perception Framework for Walking Mode Prediction
IEEE Sensors Journal · 15.07.2026 · 26 (14) · 21750-21760
DOI: 10.1109/jsen.2026.3698401 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Sensors Journal, erschienen am 15.07.2026. Verfasst von Chaoyu Song und 3 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Open Fund Program of the National and Local Co-Build Humanoid Robotics Innovation Center.
Abstract (Originalsprache)
Walking mode prediction is a critical intelligent capability for powered lower-limb prostheses (PLLPs) to achieve smooth transitions across different terrains. Inspired by human visual anticipation, this article proposes CFP-WMP, a hierarchical coarse-to-fine environmental perception framework. The system operates through four integrated modules. Initially, a coarse perception module identifies potential transition-critical terrains, triggering a fine perception module for high-precision local mapping only when a target enters the projected heading area; this selective activation ensures computational efficiency via a low-power standby state. Concurrently, a foothold prediction module estimates subsequent contact points based on historical trajectories. Finally, a locomotion mode prediction module fuses terrain geometry with these predicted footholds to anticipate gait transitions (e.g., from level ground to stair ascent), providing essential feed-forward guidance for gait control. Deployed on an NVIDIA Jetson AGX Orin, the framework achieved an average prediction accuracy of 98.92% across diverse lighting and weather conditions. With a processing latency of 92 ms/frame, the proposed framework demonstrates significant potential for real-time control, ensuring safer and more natural gait transitions for amputees in complex environments.
Förderung
- National Natural Science Foundation of China
- Open Fund Program of the National and Local Co-Build Humanoid Robotics Innovation Center
Zitieren
APA
Song, C., Chen, C., Zhang, X., & Xu, J. (2026). CFP-WMP: A Coarse-to-Fine Environmental Perception Framework for Walking Mode Prediction. IEEE Sensors Journal, 26(14), 21750-21760. https://doi.org/10.1109/jsen.2026.3698401
Vancouver
Song C, Chen C, Zhang X, Xu J. CFP-WMP: A Coarse-to-Fine Environmental Perception Framework for Walking Mode Prediction. IEEE Sensors Journal. 2026;26(14):21750-21760. doi:10.1109/jsen.2026.3698401
BibTeX
@article{Song2026,
author = {Song, Chaoyu and Chen, Chao and Zhang, Xiaoxu and Xu, Jian},
title = {CFP-WMP: A Coarse-to-Fine Environmental Perception Framework for Walking Mode Prediction},
journal = {IEEE Sensors Journal},
year = {2026},
volume = {26},
number = {14},
pages = {21750--21760},
doi = {10.1109/jsen.2026.3698401}
}
EUVIMED