Interactive Robot Perception & Learning

Projects

NEUGRASPNET: Learning Any-View 6DoF Robotic Grasping in Cluttered Scenes via Neural Surface Rendering
ACTPERMOMA: Active-Perceptive Motion Generation for Mobile Manipulation
SE(3)-DiffusionFields: Learning smooth cost functions for joint grasp and motion optimization through diffusion
Safe reinforcement learning of dynamic high-dimensional robotic tasks: navigation, manipulation, interaction
Robot Learning of Mobile Manipulation using Reachability Behavior Priors
Regularized Deep Signed Distance Field for Reactive Motion Generation
Graph-based Reinforcement Learning meets Mixed Integer Programs
Learning Implicit Priors for Motion Optimization
Active Exploration for Robotic Manipulation
Orientation Attentive Robotic Grasp Synthesis with Augmented Grasp Map Representation