ReMotion-AI: A Human-in-the-Loop Adaptive Control Framework for Compensation-Aware Rehabilitation Gaming
Shabnam Sadeghi Esfahlani, Jason D. Taylor
Anglia Ruskin University
阅读操作
确认中在文库中上传 PDF 后可生成中文音频讲解。
摘要与影响
Rehabilitation systems often face barriers to adoption due to specialised hardware requirements and limited cross-platform delivery. We present ReMotion-AI, a gamified rehabilitation platform that has transitioned from a depth-sensor desktop application to a monocular, browser-deployed solution. The system uses a server-side inference pipeline that combines Human Mesh Recovery (HMR)-based full-body 3D pose and mesh recovery, SmoothNet for temporal stabilisation, and MediaPipe Hands for lightweight real-time hand-state estimation. A key engineering objective is backward compatibility and the inference output is mapped to a standardised joint format.To support safe and clinically interpretable adaptation, an Impairment and Compensation Adaptation Module (ICAM) is introduced based on interpretable rules and calibrated scoring. ICAM estimates movement capability, including range of motion (ROM) and smoothness and detects compensatory strategies such as trunk substitution, shoulder elevation and lower-limb compensations including hip hiking and altered gait patterns. The system is formulated as a human-in-the-loop adaptive control framework, in which ICAM acts as a supervisory controller with bounded adjustment.Preliminary technical validation showed a median end-to-end latency of 180 ms, a 25–30% reduction in upper-limb trajectory jitter after temporal smoothing and stable supervisory adaptation with no abrupt difficulty changes observed under noisy monocular measurements. The proposed framework demonstrates that webcam-based pose estimation, temporal smoothing and rule-based compensation monitoring can be integrated into a browser-deployed platform.
逐年被引趋势
暂无年度引用数据
关键指标
同类平均 = 1
同领域 · 同年份 · 同类型
Google Scholar 与 OpenAlex 的被引统计范围不同,数值存在差异属正常。
AI 辅助阅读
依据:摘要
可就本文提问;依据不足时会说明。
学术脉络
学科主题
生物医学Stroke Rehabilitation and Recovery
Prosthetics and Rehabilitation Robotics · Human Pose and Action Recognition