A brain–computer-interface driven forearm exoskeleton with adaptive neuroregulation-based feedback for stroke rehabilitation
Brad Wu
- 发表年份
- 2025
- 引用次数
- 3
摘要
Exoskeleton robots hold great promise for stroke rehabilitation with upper limb impairments. However, conventional motion-based exoskeletons often fall short by lacking brain-driven control, limiting user engagement, and providing insufficient support for motor learning and neuroplasticity. These systems typically operate in a passive or pre-programmed manner, offering little adaptability to the user’s cognitive state or real-time intentions—factors that are critical for effective neurorehabilitation. This project introduces a next-generation neuroregulation-based rehabilitation platform. At its core is a brain–computer interface (BCI) that detects motor intention and identifies Alpha Wave Suppression — a neural marker of active mental engagement — enabling the exoskeleton to respond dynamically to the user’s cognitive state. A custom-designed, ultra-low-cost ( < $100) EEG acquisition module, paired with the customized BrainFormer deep learning model, enables highly accurate neural signal decoding, achieving > 92% classification accuracy for steady-state visual evoked potentials (SSVEP) and 100% two-state recognition accuracy for Alpha Wave Suppression. These metrics affirm the system’s reliability in delivering precise and intention-driven rehabilitation. The system integrates a lightweight ( < 1.5 kg), five-degree-of-freedom (5-DOF) forearm exoskeleton with real-time haptic vibration feedback, forming a closed-loop and user-adaptive architecture. Unlike traditional exoskeletons that rely solely on mechanical motion cues, this neuroadaptive design leverages real-time brain signals to personalize movement assistance and encourage continuous user involvement. This advanced exoskeleton provides a more efficient and responsive solution than conventional motion-based systems. It not only enhances rehabilitation outcomes but also offers a scalable, cost-effective tool to accelerate motor recovery and foster neuroplasticity in individuals recovering from subacute and chronic stroke.
关键词
相关论文
Artificial intelligence: a modern approach
1995
Self-Organizing Maps
Teuvo Kohonen
1995
Vision meets robotics: The KITTI dataset
Andreas Geiger, Philip Lenz, Christoph Stiller 等 4 位作者
2013
Probabilistic robotics
Sebastian Thrun
2002