Tracking and Catching Balls with Motion Capture and a Robot
Session Number
3
Advisor(s)
Pranav Audhut Bhounsule, University of Illinois Chicago
Location
B116
Discipline
Computer Science
Start Date
15-4-2026 2:15 PM
End Date
15-3-2026 3:00 PM
Abstract
Catching moving objects represents a fundamental challenge in robotics, requiring the integration of real-time vision, data processing, trajectory prediction, and dynamic control. Numerous previous approaches use onboard vision systems, but this paper instead presents a framework that uses an external motion capture system to achieve highly precise tracking and interception of a moving ball. An OptiTrack motion capture system provides real-time positional data of thrown balls, which feeds into trajectory prediction algorithms that estimate the ball's flight path by using a Kalman filter. A bipedal robot then processes the predicted trajectory to position its arm for successful interception. This approach provides insights into the relative importance of trajectory prediction and arm control components in dynamic catching tasks. This demonstrates the viability of motion capture-based systems for applications involving robots catching moving objects. The results show that external motion capture can provide a robust foundation for real-time trajectory prediction and interception, with implications for human-robot collaboration, assistive robotics, and interactive entertainment applications where precise dynamic object manipulation is required.
Tracking and Catching Balls with Motion Capture and a Robot
B116
Catching moving objects represents a fundamental challenge in robotics, requiring the integration of real-time vision, data processing, trajectory prediction, and dynamic control. Numerous previous approaches use onboard vision systems, but this paper instead presents a framework that uses an external motion capture system to achieve highly precise tracking and interception of a moving ball. An OptiTrack motion capture system provides real-time positional data of thrown balls, which feeds into trajectory prediction algorithms that estimate the ball's flight path by using a Kalman filter. A bipedal robot then processes the predicted trajectory to position its arm for successful interception. This approach provides insights into the relative importance of trajectory prediction and arm control components in dynamic catching tasks. This demonstrates the viability of motion capture-based systems for applications involving robots catching moving objects. The results show that external motion capture can provide a robust foundation for real-time trajectory prediction and interception, with implications for human-robot collaboration, assistive robotics, and interactive entertainment applications where precise dynamic object manipulation is required.