In my advanced robotics class in my final semester of school, everyone was tasked with implementing a completely novel project within the last three weeks of the semester. The professor suggested I do "tactile based scanning" using the campus's UR5e robotic arm. There are existing projects and papers discussing scanning using robotic arms, but most implementations use cameras.
So why would you ever want to use a "feeler" to "scan" a surface? You wouldn't! You just can't get as many datapoints as quickly as you could with a vision system. But we needed something novel for the project.