For much of the industrial age and continuing to today, the goal of creating flexible, mixed-use robotics has been the holy grail of the business world. Of course, multiple industries have long relied on purpose-built, task-specific robotics for some time, but they have natural limitations. One of those limitations is that most of today’s robotics solutions need very specific working conditions to function.
That’s one of the reasons that retail giant Walmart recently ended a plan to introduce inventory checking robots to its stores. After a 500 store trial, they found that it was simpler – and cheaper – to just let humans do the work instead. In other words, many potential robotics applications are waiting on robots to become more human, or at least more closely mimic the capabilities of humans.
And that’s why a recent announcement by researchers from Cornell University might turn out to be a watershed moment in robotics. It involves a breakthrough in sensor design that might soon make it possible for robotics developers to create human-like machines to handle a variety of everyday tasks. Here’s an overview of what the announcement entails and what it might mean.
The Human Touch
On November 13th, materials and robotics researchers at Cornell University in upstate New York published a paper in the journal Science entitled, “Stretchable Distributed Fiber-Optic Sensors“. In it, they detailed their work on a new breed of light sensors that use low-cost dye and off-the-shelf LEDs to create the machine equivalent of a human-like skin. Just like human skin, it can stretch, bend, and deform as needed.
More importantly, however, it can also detect and measure those movements with a high degree of accuracy. In other words, it’s a robotic skin that can mimic our sense of touch. That raises the possibility of creating fully-articulated soft robotic interfaces that can interact with the world around them in the same way humans do.
How the Sensors Work
The breakthrough builds on other research that came from the university in recent years. Cornell’s Organic Robotics Lab is known for a string of similar work, including a stretchable lace sensor system as well as another that used foam to help robotics systems to sense the environment around them. In both cases, however, the materials could only detect certain kinds of environmental conditions at any given time.
To improve on them, the newest version uses silica-based distributed fiber-optic sensors and two polyurethane elastomeric cores working in parallel. One of the cores contains a light-absorbing dye and the other is transparent. And each connects to an RGB sensor that can detect minute changes in light transmission passing through its respective core.
That makes it possible for the sensors to detect multiple types of deformations at the same time. And when coupled with a custom mathematical model, it can localize the effects and even measure their intensity. In other words, the sensor system knows exactly where it’s being touched, and how much pressure is being applied.
The Potential Applications
It doesn’t take much of an imagination to think of the potential uses for this new sensor material. But where it has the most potential is in automating existing workflows in a less disruptive way. For example, in a laboratory environment, one could automate the use of a standard BMG Labtech microplate reader or facilitate the use of glassware pipettes, both of which were designed for human operators.
Using today’s robotics and automation technologies, such automation often requires the wholesale replacement of equipment – forcing a decision between a human-centric or robot-centric environment. Using touch-sensitive soft robotics interfaces, it will become possible to have the best of both worlds: automation where applicable working right alongside human staff.
The Bottom Line
In the near-term, the sensor system’s designers expect that it will see use in augmented reality and within physical therapy and sports medicine. Those are fields that have relied on camera-based motion tracking for some time, but that could use the new sensors to create more precise and useful equipment.
But if those real-world applications prove successful, it shouldn’t take long to see the sensor system start to become a core part of next-generation soft robotics systems. It’s obvious that there’s plenty of appetite and applications for such robotics systems, and this could be the breakthrough that makes many others possible.