Search papers, labs, and topics across Lattice.
5
0
6
Achieving zero friction violations and the best composite score in dexterous grasp generation could redefine standards for robotic manipulation.
Variational neural belief parameterizations can drastically enhance grasping success rates and reduce planning time under multimodal uncertainties, outperforming traditional methods by an order of magnitude.
Achieve correct-by-construction control for autonomous systems by encoding Temporal Behavior Trees with ternary logic, allowing for reasoning about partial satisfaction of temporal specifications.
Imperfect robot demonstrations don't have to sink your RL: this method uses Temporal Behavior Trees to automatically repair trajectories, leading to more efficient and reliable learning.
Key contribution not extracted.
