Touch Blind: The Quest for Grippers That Feel

Touch blind
In an effort to enable robots to work faster and more accurately, technology is fast about loading them up with all sorts of new sensor gear, and they’re doing more varied work because of it. Lights, camera, action! Most of their new duties revolve around vision systems.

However, if robots are to take on ever more jobs from humans, they will inevitably benefit from adding ever more human-like senses to their limited repertoire of sight and sound. One of the most elusive is the human sense of touch. Robots cannot feel what they touch; they are touch blind.

For most industrial robots the sense of touch—being able to feel what they are working on—would be a luxury without purpose. Programmed to do repetitious jobs, most times robots need only a good vision system.

But that could change. Liberating robots from touch blindness would allow for the grippers on industrial robots to handle, understand, and give feedback in dealing with soft, fragile and deformable objects, which, in turn, could open up new avenues of manufacturing.

In their 2019 review of haptics or active touch in robotics, the authors of Active Haptic Perception in Robots: A Review (Frontiers of Neurorobotics) show lots of lab work in haptics for robots going on but not much in the real world of marketable products.

For the importance of haptics in robot manipulation, they show three broad areas of need. As they put it: “Haptics provides useful information under a wider set of circumstances than is sometimes appreciated.”

1. Obviously, haptics can be the only available signal, such as in poorly lit environments.
2. Even when other senses are available, haptics provides direct information about properties of objects that are poorly (or at least very indirectly) specified by other senses. This is because haptics, uniquely among human senses, involves physical interaction with objects, allowing properties such as friction and compliance to be sensed.
3. Haptics contributes to sensorimotor processing even when it provides redundant information to other senses. Our perception of the size of an object held in our hand, for example, is derived by integrating information from vision and haptics…we constantly make use of haptic information to optimize perception (though we may not be aware of it

That being said, although MIT’s STAG haptic glove (still in R&D) with its 500-plus sensors across the palm and fingers (see photo) will be a great advance, especially for prosthetics, there’s not much available for the plain old—but ubiquitous—two-fingered industrial gripper.

An exception is a very new and very interesting product lineup from Contactile, a small, newbie Australian (Sydney) company founded in 2019 that shouts out from its homepage: Advancing Industries Through an Intelligent Sense of Touch. “We are a sensor + haptics company specializing in tactile sensors for real-world robotic applications.”

Not yet a pinball wizard
Dr Heba Khamis, one of the three co-founders of Contactile, says of her company’s tech: “Most touch-based sensors that are used in robotics are either 3D force/torque sensors or pressure sensor arrays. These sensors can tell how hard the robot is gripping, but can’t determine how hard the robot needs to be gripping.”

Khamis, and her two Contactile co-founders, Benjamin Xia and Dr Stephen Redmond collaborated on a research project at the University of New South Wales (Sydney) that resulted in the development of Contactile’s technology.

Their sensor technology won the Advanced Manufacturing Award at the Australian Technologies Competition 2020. Well-deserved and commendable she remarks about the award, but even better: “Our sensor enables robots to perform as dexterously as humans do.”