On the generality and application of mason’s voting theorem to center of mass estimation for pure translational motion
Published in IEEE Transactions on Robotics, 2024
🤖 Mason's Voting Theorem for Center of Mass Estimation
Extending Mason Voting Theorem to robust center of mass estimation under uncertainty in friction and object geometry
📋 Overview
This work extends Mason’s Voting Theorem to object center of mass estimation during pure translational motion, addressing the challenge of estimating center of mass when accurate information on friction and object shape is unavailable. By leveraging a position-controlled robot arm combined with vision sensing, the method provides a robust and generalizable approach to center of mass estimation.
Key Contributions
- Theoretical generalization of Mason Voting Theorem for center of mass estimation
- Robust estimation methodology independent of friction model and object geometry
- Integration of vision-based sensing with robotic manipulation
- Position-controlled pushing experiments validating the approach
- Practical applicability for robotic object manipulation tasks
🎨 Methodology Overview
The proposed approach demonstrates how Mason Voting Theorem can be generalized to estimate the center of mass of objects under pure translational motion. By applying pushing motions at various contact points and observing the resulting object trajectories using vision sensing, the method aggregates these "votes" to accurately determine the center of mass, even without precise knowledge of friction coefficients or object geometry. This voting-based approach provides robustness and generalizability across different object shapes and surface properties.
🎥 Demonstration Videos
Video 1
Center of Mass Estimation with Iterative Pushing demonstration
Video 2
Center of Mass Estimation with Iterative Pushing demonstration
Video 3
Center of Mass Estimation with Iterative Pushing demonstration
Video 4
Center of Mass Estimation with Iterative Pushing demonstration
Video 5
Center of Mass Estimation with Iterative Pushing demonstration
Video 6
Center of Mass Estimation with Iterative Pushing demonstration
Description: This video demonstrates how the object center of mass (CoM) is estimated through iterative pusher-object interactions. The pushing actions are selected based on QP-EVT (Quadratic Program - Entropy and Variance Trade-off) based sampling strategy, which efficiently explores the object's contact surface to gather informative pushing "votes". The system uses vision sensing to track object motion and aggregates the voting results to accurately estimate the CoM without requiring prior knowledge of friction properties or object geometry.
🎬 Object Translation Videos
Translation 1
Pure translational motion demonstration
Translation 2
Pure translational motion demonstration
Translation 3
Pure translational motion demonstration
Translation 4
Pure translational motion demonstration
Translation 5
Pure translational motion demonstrationn
Translation 6
Pure translational motion demonstration
Description: These videos showcase pure translational motion of objects during pushing interactions. The contact configurations between the pusher and the object were sampled based on modified Zero Moment Two Edge Pushing (ZMTEP) method.
🎤 IROS 2024 Oral Presentation
Oral Presentation at IROS 2024
This is the oral presentation of "On the generality and application of mason's voting theorem to center of mass estimation for pure translational motion" presented at the 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS 2024).
📄 Publication Details
Authors: Ziyan Gao, Alp Elibol, Nak Young Chong
Venue: IEEE Transactions on Robotics
Volume: 40
Pages: 2656-2671
Publication Date: 2024-04-22
DOI: Read on IEEE Xplore
Recommended citation: Gao, Z., Elibol, A., & Chong, N. Y. (2024). On the generality and application of mason's voting theorem to center of mass estimation for pure translational motion. IEEE Transactions on Robotics, 40, 2656-2671.
Download Paper