Tactile Sensors for Manufacturing: How Robots 'Feel'

Watch this video to see Howie Choset, Lu Li, and Victoria Webster-Wood — from the Manufacturing Futures Institute — explain their work. It creates specialized sensors that allow robots to "feel" the world around them.



Transcript

00:00:10 Howie: One of the things that make us,  or all animals, successful are sensors.   But sensing is just a hard problem.  And we see with our eyes, obviously,   but we also feel. We can also sense  force. And building those kinds of   sensors has proven to be very difficult for  industry. Lu Li, he sort   of had this little hobby on the side for  building four sensors on this one project. Lu: And so the goal for this project is to  build smart, intelligent tactile sensors to   get robots sense of touch so that we can use  the sensor for challenging applications such   as medical, assembly, disassembly,  and manufacturing applications. Howie: When you go and, say, grab that doorknob or  put the key in the keyhole, you are not perfectly   positioning your joints just right, you're  approximating it, but you're feeling your way  

00:01:06 through. And to do that, you need force feedback,  force sensing. The problem with the force sensors   that exist today are they're very expensive,  and they're finicky, and they break easily. Vickie: Effective, low-cost tactile sensing  is an ongoing challenge in manufacturing and   recycling robots. Without the sense of touch,  these robots rely on expensive camera systems   with high-power computers to know where objects  are in their environment. In contrast, animals and   humans are able to explore much of our world via  our sense of touch. So part of our ongoing work is   to try to bring some of these capabilities  from human touch into robotic sensors. Lu: We are able to use magnetic field  to indirectly measure tactile and   force information. So this is novel because  magnetic field is one of the special signal   that we can take advantages to convert contact  information directly into electrical signal. And  

00:02:06 this is super valuable since we now can understand  what is the magnitude and direction of contact. Vickie: In this work, we've taken inspiration from  human fingers in order to create a low-cost and   robust tactile sensor for robotics. We've mimicked  the different layers of tissue that you see in the   human finger, as well as fingerprint ridges on the  surface of the sensor. And together these features   allow our sensor to sense both pressure  and shear as well as vibration signals   if we pull the sensor across the  surface, and so this lets us do   things like surface classification and material  classification. Because the sensor is low-cost,   it can be easily integrated and replaced as  needed in demanding manufacturing environments. Lu: Tactile sensors are integrated into  a parallel gripper, which in our case   is a grasper that can touch and grasp object to  apply assembly and also disassembly operations.

00:03:06 Vickie: Many of the ongoing challenges in  incorporating this type of tactile sensing   into manufacturing robots are really  around cost and robustness. You need   the sensors to be low enough cost that they  can be readily deployed on fleets of robots,   and they need to be robust enough that you can  trust the data and you're not having to replace   them frequently. And so this work really starts  to address many of these challenges by bringing   the cost down in such a way that these sensors  are more accessible for manufacturing partners. Howie: And this sensor, you can take  a hammer and mallet it if you want,   you can rub a feather over it and it  will detect the force interactions that   are having. And that's necessary when you're  performing a variety of tasks such as assembly. Lu: In the Center for Manufacturing  Future Institute we build a rockstar team  

00:03:54 combining roboticist, biological inspired  group, and also engineers to build a unique   sensor that can help promoting better intelligence  in medical, manufacturing, and service industry.