Soft-assist robotic wearables benefit from a rapid design tool
Scientists have created a new design and manufacturing tool for flexible pneumatic actuators with built-in sensing, which can power personalized healthcare, smart homes and gaming.
Soft pneumatic actuators may not be a phrase that comes up in everyday conversation, but more than likely you have benefited from their usefulness. The devices use compressed air to power movement, and with sensing capabilities they have proven to be an essential component in a variety of applications such as wearable assistive devices, robotics and rehabilitation technology .
But there’s a bit of a bottleneck in creating small, dynamic devices that have advantages like high response rates and power-to-input ratios. They require a manual design and manufacturing pipeline, which results in many trial and error cycles to test and see if the designs will work.
Scientists at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) have designed a scalable pipeline to computer-design and digitally manufacture flexible pneumatic actuators called “PneuAct”.
PneuAct uses a machine knitting process – not much different from knitting with your grandmother’s plastic needle – but this machine works on its own.
A human designer simply specifies the point and sensor design patterns in software to program the movement of the actuator, and it can then be simulated before printing. The textile part is made by the knitting machine, which can be attached to a cheap ready-made rubber silicone tube to complete the actuator.
The knitted actuator incorporates a conductive thread for sensing, allowing actuators to “sense” what they are touching. The team concocted several prototypes covering a support glove, a soft hand, an interactive robot and a pneumatically walking quadruped. Their devices were wrapped in a soft, yellow fabric that made them look a bit like banana fingers.
Although there has been a lot of movement in the hardware development of soft pneumatic actuators over the years – a 2019 prototype of a collaborative robot used such actuators to replicate a human-like grip in its hands – the design tools did not improve with such speed.
Older processes typically used polymers and molding, but scientists have used a combination of elastic and sensing points (with a conductive thread) that allow the devices to be programmed to flex when inflated and the ability to integrate a real-world feedback.
For example, the team used the actuators to build a robot that senses when it is specifically touched by human hands and reacts to that touch.
The team glove can be worn by a human to complement finger muscle movement, minimizing the amount of muscle activity required to complete tasks and movements. This could have a lot of potential for people with injuries, limited mobility or other finger trauma.
The process can also be used to make an exoskeleton (wearable, computer-controlled robotic units that complement human movement and restore locomotion and movement); and for this purpose they have made a sleeve which can help the wearers to bend the elbow, the knee or other parts of the body.
Yiyue Luo, a PhD student at MIT CSAIL and lead author of a new paper on the research, says, “Digital machine knitting, which is a very common manufacturing method in today’s textile industry, enables” to print” a pattern in one go, which makes it much more scalable.
“Soft pneumatic actuators are inherently compliant and flexible, and combined with smart materials, they have become a necessary force in many robots and assistive technologies – and rapid manufacturing, with our design tool, can hopefully , increase ease and ubiquity.”
Giving meaning to sensors
One type of sensing the team incorporated was called “resistive pressure sensing,” where the actuator “sends” the pressure. When making a robotic gripper, it would try to grab something, and the pressure sensor would detect the force applied to the object, then it would try to see if the grip was successful or not.
The other type is “capacitive sensing”, where the sensor discerns certain information about the materials the actuator comes into contact with.
The actuators are sturdy – no wires were damaged in the process – a limitation of the system is that they were limited to tube shaped actuators as they are very easy to buy off the shelf. A logical next step is to explore actuators of different shapes, to avoid being constrained by this unique structure.
Another extension that scientists will explore is extending the tool to incorporate task-based optimization-based design, where users can specify target poses and optimal point patterns that can be automatically synthesized.
Andrew Spielberg, a postdoctoral researcher in materials science and mechanical engineering at Harvard University, another author of the paper, says, “Our software tool is fast, easy to use, and it accurately previews user designs, allowing them to quickly iterate virtually while only needing to manufacture once. But this process still requires some trial and error on the part of humans.
“Can a computer reason about how textiles need to be physically programmed into actuators to enable rich, sensing-driven behavior? This is the next frontier.
Luo authored the paper alongside former MIT CSAIL doctoral student Kui Wu, Spielberg, MIT CSAIL mechanical engineer Michael Foshey, and MIT professors Tomas Palacios, Daniela Rus, and Wojciech Matusik.