Deep reinforcement learning with
robot touch
Collaboration with DeepMind
Robot touch of virtual stimuli from ultrasonic haptic displays
Collaboration with Ultraleap
In-hand manipulation with a
fully-actuated tactile robot hand
Collaboration with Shadow Robotics
A biomimetic forebrain for robot touch
Leverhulme Research Leadership Award

Deep learning for controlled robot touch
Leverhulme Research Leadership Award
Tactile robot hands
Leverhulme Research Leadership Award
Control methods for robot touch
PhD projects
(collaboration with Dave Barton)
Biomimetic slip detection and correction
PhD project
Tactile superresolution sensing
EPSRC first grant

Research Themes

Robot sensors, hands & systems with a sense of touch

The BRL Tactile Robotics group designs and builds novel dexterous tactile robots from 3D-printed fingertips/hands to entire integrated robot systems.

Our technology is based on the TacTip family of tactile fingertips:
Soft – Multi-material 3D-printed
Biomimetic – like the human sense of touch
Optical – based on using an inner tiny camera
Tactile – measures the intricacies of contact

Artificial Intelligence for robot dexterity

We develop artificial intelligence that enables our tactile robots to perceive, explore, manipulate and handle complex objects and environments.

Our group is at the forefront of applying deep learning to robot dexterity. Neural networks are revolutionizing computer and robot vision. We aim to transfer those advances to robot touch.

Neuroscience of human & animal perceptual & action

Our group’s work is also inspired by the neuroscience and psychology of perception and action. Human dexterity eclipses that of all other animals and is fundamental to human intelligence.

Our interests span from sensory processing and decision making to models of perception and action selection.