Poor posture and incorrect movement patterns are common in both exercise (e.g., squats, lifting) and daily behaviors (e.g., prolonged sitting, screen usages). These patterns can gradually lead to discomfort, pain, or even injury.
This project investigates: How can we build an embodied, closed-loop coaching system that detects risky posture/movement and delivers actionable, low-burden feedback in real-world settings?
Students are encouraged to propose and prototype their own solution. Possible approaches include (but are not limited to) smartphone sensing, wearables, environmental sensing, computer vision, interactive feedback, and data-driven modeling. The project is problem-first and solution-open, focusing on health support and prevention, not medical diagnosis.
Expected deliverables (by the end of the project):
1. Conduct a brief literature review and refine the broad topic into a focused, testable research problem.
2. Define measurable “risk posture/movement” indicators, specify target errors, and establish evaluation metrics.
3. Build a Minimum Viable Prototype (MVP) that implements the proposed approach.
4. Validate usefulness through a pilot evaluation (e.g., a small user study or dataset-based validation).
5. Produce a final report and presentation/poster summarizing the method, results, design implications, and limitations.