2020-2021 PPRT Project

Computer Vision Predicts Force During Lifting

Research Trainee: Guoyang Zhou, PhD Student, School of Industrial Engineering, Purdue University

Co-PI: Denny Yu, PhD, Assistant Professor, School of Industrial Engineering, Purdue University

Co-PI: Vaneet Aggarwal, PhD, Associate Professor, School of Industrial Engineering, Purdue University

Guoyang Zhou, PhD Student

Denny Yu, PhD

Vaneet Aggarwal, PhD

Repetitive and overexertion in lifting are one of the most significant causes of Musculoskeletal disorders (MSDs) at workplaces. To minimize these MSDs due to overexertion in lifting, early detection of lifting risk is needed. However, current techniques of evaluating the lifting risk have limitations. Sensor-based (e.g., surface muscle activity sensors and motion sensors) techniques are highly intrusive to worker, while aforementioned assessment-based approaches, such as RULA and NIOSH lifting equation, require variables that are hard to attained in real-time, such as the weight of the object being lifted. In this study, we purposed a non-intrusive and real-time based technique to predict lifting risk to overcame to limitations of existing techniques. The purposed technique involves using computer vision techniques to monitor people’s facial expression, body posture and motion. Photoplethysmogram (PPG) features will be integrated into the purposed technique to reduce the uncertainties if needed.