Coordinated dynamic manipulation: enhancing task completion rate

Coordinated dynamic manipulation: enhancing task completion rate
Multi-arm control / Hierarchical reinforcement learning with motor primitives

Reaching and grasping in 3D

reaching and grasping in 3D
Twisting motion with oblique muscle model for solving tasks in 3D

Continuous deformation mode reconstruction from video data

robotic-arm-overlaid-reconstruction
Pose tracking with Optical Flow/ Physics-informed method to estimate all six mode of strains

Multi-arm coordination for foraging

Whole-arm manipulation: budge and pitch
Hierarchical decomposition / Multi-agent reinforcment learning / Local sensory feedback control

Dynamic whole-arm manipulation

Whole-arm manipulation: budge and pitch
Transfer with motion primitives / Reinforcement learning

Dynamic primitives of an octopus arm

Dynamic primitives of an octopus arm
Motion decomposition / Learning from octopus bend-propagation with imitation learning

Real-time control of an octopus arm (NNES)

NNES is 200 times faster than ES
Octopus muscular model / Compact representations of continuous actions /
Energy shaping control / Unsupervised learning of reaching
Enter Blog