Inexpensive 'GPS' for Flexible Medical Robots
Roboticists at UC San Diego have developed an affordable, easy-to-use system to track the location of flexible surgical robots inside the human body. The system performs as well as current state-of-the-art methods, but is much less expensive. The researchers embedded a magnet in the tip of a flexible robot that can be used in delicate places inside the body, like arterial passages in the brain. The researchers then used existing magnet localization methods, which work a lot like GPS, to develop a computer model that predicts the robot’s location.
Transcript
00:00:00 so the type of work we're doing is important for these types of continuum medical robots mainly because these robots are great for these types of environments inside the body these highly constrained environments and these robots are able to conform to these types of cavities etc and they're inherently safer much more compliant but it becomes a lot harder to actually
00:00:22 track their location and their shape inside the body and so if we are able to do that more easily then that would be a great benefit to both the patients and the surgeons we did work with a growing robot which is robot made of a very thin ripstop nylon that we invert inside itself almost like a sock and then pressurized with the fluid which causes the robot to
00:00:48 grow we think this robot has a lot of potential for use in medical applications because it's inflatable it's soft and it transmits very little force to its surroundings so it can be used in delicate places like arterial passageways in the brain so we looked at using magnets instead to track the tip of the robot so that you could know where the robot is without having to
00:01:12 constantly image under fluoroscopy it's now to track the tip of the robot we should have to be able to track the magnet we have an idea of where or of how strong the magnetic field should be around the magnet and so if we have a bunch of sensors around the magnet that measure the magnetic field strength we can determine where the magnet is we trained a neural network to learn the
00:01:32 difference between what our sensors were reading and what our model said this such as should be reading and then together using that physics based model and our neural network we improved our localization accuracy and tracking the tip of these growing robots yeah all together our four sensors are magnet and our microcontroller it's about $100 system I guess ideally we are hoping
00:01:59 that we can help develop these types of growing robot technology push it forward so that we can actually try to test it in a clinical setting move it forward and actually be translated into clinical use