I am a fourth-year PhD student in Computer Science at the
University of Colorado Boulder, advised by
Alessandro Roncone in the HIRO Group.
My research develops learning-based motion generation models, deployed on real
manipulator hardware.
02
Research
02.1
Dance2Hesitate: A Multi-Modal Dataset of Dancer-Taught Hesitancy for Understandable Robot Motion
Designing Transparent and Understandable Robots (D-TUR) Workshop at HRI · 2026
An open-source, multi-modal dataset of dancer-generated hesitant motion: 66
kinesthetic-teaching demonstrations on a Franka Emika Panda reaching toward a Jenga
tower at three graded hesitancy levels, plus synchronized RGB-D motion capture of 84
upper-limb and 70 whole-body dancer trajectories.
Employing Laban Shape for Generating Emotionally and Functionally Expressive Trajectories in Robotic Manipulators
IEEE International Conference on Robot and Human Interactive Communication (RO-MAN) · 2025
Uses Laban notation to give manipulators functional expressivity: two new variants of
hesitant motion, spoke-like and arc-like, that participants associate with reduced robot
competency, and Happy, Sad, Shy, and Angry trajectories enhanced with Laban Shape that
read as distinct emotions on the valence-arousal-dominance spectrum.
BBoE: Leveraging a Bundle of Edges for Kinodynamic Bidirectional Motion Planning
IEEE Robotics and Automation Letters (RA-L) · 2026 · Accepted
A bidirectional, kinodynamic, sampling-based planner that reuses a precomputed bundle
of robot motions to converge quickly in cluttered environments. Across four systems
spanning 3D to 6D state spaces, BBoE is the only planner to reach ≥96% success on
every system, with the lowest planning time on three of the four.
Full planning runUnicycle, easyUnicycle, mediumUnicycle, hardUnicycle, very hardQuadrotorTreaded vehicleCar with trailer
