|
Double-T DINER: Deploying Improvements in New Environments via Refined Tool Transfer
Henry Tsai*, Leo Huang*, Denae Cantrell*, Narek Harutyunyan, Hunter Tsai, Maggie Dong, David Jo, Claire Tomlin, Shankar Sastry
CoRL 2026 Late-Breaking Results & Sim2Real2Field Workshop; under submission
website
A refiner agent hill-climbs a Spot robot agent’s toolset in simulation, and the improved tools carry over to 220 real-world field trials.
|
|
Learning to Grasp Anything by Playing with Random Toys
Dantong Niu*, Yuvan Sharma*, Baifeng Shi*, Rachel Ding, Matteo Gioia, Haoru Xue, Henry Tsai, Konstantinos Kallidromitis, Anirudh Pai, Shankar Sastry, Trevor Darrell, Jitendra Malik, Roei Herzig
ICLR, 2026
arxiv / website
Training on randomized toys built from four shape primitives lets a robot grasp 64 real-world objects zero-shot, with 67% success on YCB.
|
|
Autonomous Airdrop
EECS 106A final project, 2024
website / code
We built a fixed-wing drone that flies a planned route and drops a payload on a target by itself, with the first autonomous hit in December 2024. I built the HSV tuner and the image detection pipeline.
|
|
FTC Team #16236 Juice
Sacramento-area education nonprofit, 2019 — present
website
I founded Juice, a community FIRST Tech Challenge team, which I captained to win the 2020 NorCal Regional Championship in our rookie season and where I currently serve as lead mentor (two playoff runs at the 2023 and 2026 FIRST Championships).
|
|