Electronic leather for robots

History

2025: Product Announcement

Innovative electronic "leather" for robots, developed in collaboration between the University of Cambridge and University College London, presented in June 2025. The new synthetic coating provides robotic systems with the ability to perceive physical exposure from the outside environment - including temperature, pressure, multiple touches and damage - through a single sensory material, bringing their sensory capabilities closer to human touch.

According to TechXplore, the developed technology is a soft conductive material that can take complex forms and adapt to various tasks. Unlike traditional sensor systems that use separate sensors to transmit different types of sensations, the new solution functions as a single sensory surface.

𝓲

Фото: techxplore.com

A key feature of electronic skin is its internal architecture, containing more than 860 thousand microscopic channels for signaling. This structure allows the material to recognize many kinds of interactions with the environment at the same time.

Artificial skin is able to distinguish between light touches, high and low temperatures, damage to sharp objects and pressure on several different points on the surface. The researchers applied machine learning technologies to identify the most important data channels for each type of interaction.

Optimization of the system through machine learning algorithms has improved the response to various external stimuli. The neural network was trained to interpret physical signals and correctly recognize the types of effects based on the data obtained.

The material has the consistency of jelly, which allows you to melt it and cast shapes of any complexity. The low cost of production and ease of manufacture make the technology promising for mass implementation.

During the experiments, scientists created the shape of a human hand and built 32 electrodes into it, placed in the wrist area. Despite the limited number of sensor points, the system captured more than 1.7 million signals, demonstrating high perceptual density.^[1]

Notes