Vision-Enhanced Robotic Product Testing: Tracking User Interaction Trajectories
Abstract: This project focuses on automating the testing of consumer products by developing a vision-based system that captures and helps in replicating user interactions using robotic manipulators. Consumer product testing currently involves manual observation and interpretation of user behavior, which is both time-intensive and prone to inconsistencies.
The proposed solution leverages enhanced vision-based tracking to capture the relative pose and trajectory of consumer electronic devices during user interaction. Using a combination of video-based pose estimation and transformation analysis, the system generates precise motion trajectories of the product in use. These trajectories can then be replicated by an industrial robotic arm, enabling accurate and repeatable testing of the product under realistic user conditions.
This approach enhances the testing process by introducing automation, reducing variability, and improving reproducibility. Additionally, the system provides deeper insights into user behavior, contributing to product optimization and validation.
Advisor: Professor Constantinos Chamzas
Committee: Professor Greg Lewin, Professor Constantinos Chamzas, Mr. Viktor Bergman