Three members of the Interactive Robotics Group --- Claudia Pérez D'Arpino, Brad Hayes and Pem Lasota --- are presenting their work at the IEEE International Conference on Robotics and Automation (ICRA) in Singapore this week.

Przemyslaw "Pem" Lasota will be presenting his paper titled "A Multiple-Predictor Approach to Human Motion Prediction" at the main conference on May 30th at 4:50pm in the Human-Robot Interaction session. The paper presents a technique that learns from data to automatically combine several human motion prediction methods together -- so as to take advantage of each method's complementary strengths. You can read the full paper here.

Claudia Pérez D'Arpino will be presenting her paper titled "C-LEARN: Learning Geometric Constraints from Demonstrations for Multi-Step Manipulation in Shared Autonomy" at the main conference on Wednesday, May 31st at 2:50pm. The paper presents C-LEARN, a method that builds a knowledge base for reaching and grasping objects, which is then leveraged to learn multi-step tasks from a single demonstration.  C-LEARN supports tasks with multiple end effectors; reasons about SE(3) volumetric and CAD constraints; and offers a principled way to transfer skills between robots with different kinematics. You can read the paper here.

Brad Hayes will be presenting his paper titled "Interpretable Models for Fast Activity Recognition and Anomaly Explanation During Collaborative Robotics Tasks" at the main conference on Thursday, June 1st from 4:45-4:50pm in the "Human Factors 2" session. His paper presents a highly parallel technique for real-time, online activity recognition to facilitate human-robot collaboration. The presented technique achieves state-of-the-art recognition accuracy given full or partial trajectories, and is capable of producing interpretable explanations for its classification decisions to human co-workers. The full paper can be accessed here.