With rapidly increasing demands on healthcare systems, nurses typically spend between 18 and 40 percent of their time performing duties direct patient care tasks, often for many patients and with little free time. Personal care robots that brush hair could provide significant facilitate and relief.
It may seem like a radical form of “self-care,” but clever robots for tasks like shaving, washing hair and putting on makeup are nothing recent. In 2011, tech giant Panasonic developed a robot that can wash, massage and even arid hair, specifically designed to facilitate “elderly and people with reduced mobility live safely and comfortably, while reducing the burden on caregivers.”
Hair-combing robots, however, have proven to be less explored, so leading scientists at MIT’s Computer Science and Artificial Intelligence Laboratory (CSAIL) and Harvard’s Pliable Mathematics Laboratory have developed a sensor-enabled plush brush robotic arm system. The robot is equipped with a camera that helps it “see” and assess curls, so it can plan gentle and time-saving brushing.
The team’s control strategy adapts to the degree of entanglement in the fiber bundles, and the team put “RoboWig” to the test by brushing wigs with hair ranging from straight to very curly.
Although the hardware configuration of RoboWig looks futuristic and shiny, the basic hair fiber model is what makes it work. CSAIL Ph.D., Josie Hughes and her team chose to represent tangled hair as sets of intertwined double helices – think of classic DNA strands. This level of detail provided key insight into the mathematical models and control systems for manipulating plush fiber bundles, with a wide range of applications in the textile industry, animal care and other fibrous systems.
“When developing the tangled fiber model, we understand from a model perspective how the hair needs to be tangled: starting at the bottom and slowly working its way up to prevent the fibers from getting ‘stuck’,” says Hughes, lead author of the book’s article on RoboWig. “It’s something that anyone who combs their hair has learned from experience, but now we can demonstrate it with a model and use it to inform the robot.”
This task is complicated. Every head is different, and the sophisticated interplay of hair when combing can easily lead to knots. Moreover, if the wrong brushing strategy is used, the process can be very painful and harmful to the hair.
Previous research in the field of brushing has mainly focused on the mechanical, vigorous and visual properties of hair, in contrast to RoboWig’s sophisticated focus on tangles and combing.
To brush and manipulate hair, researchers added a soft-bristled sensor brush to the robot’s arm to measure the force applied during brushing. They combined this setup with something called a “closed-loop control system,” which takes feedback from the output signal and automatically performs an action without human intervention. This created the brush’s ‘force feedback’ – a control method that allows the user to feel what the device is doing – so that stroke length could be optimized to account for both potential ‘pain’ and the time required for brushing.
Initial tests – for now – preserved the human head and were instead conducted on a range of wigs with different hairstyles and types. The model provided insight into brushing behavior related to the number of tangles and how they could be effectively and efficiently brushed by selecting the appropriate brushing length. For example, with curly hair, the cost of pain would dominate, so a shorter brush length was optimal.
The team eventually wants to conduct more realistic experiments on humans to better understand the robot’s performance in relation to how they feel pain – a metric that is, of course, highly subjective, as a “two” for one person may be an “eight” for another.
“To enable robots to extend their task-solving capabilities to more complex tasks, such as combing hair, we need not only novel safe hardware, but also an understanding of the complex behavior of soft hair and tangled fibers,” says Hughes. “In addition to combing hair, the lessons from our approach can be applied to brushing textile fibers or animal fibers.”
Hughes wrote the paper with Harvard School of Engineering and Applied Science graduate students Thomas Bolton Plumb-Reyes and Nicholas Charles; Professor L. Mahadevan from the Harvard School of Engineering and Applied Sciences, Department of Physics and Organismic and Evolutionary Biology; and MIT professor and CSAIL director Daniela Rus. They presented the paper virtually at the IEEE Conference on Pliable Robotics (RoboSoft) earlier this month.
The project was supported in part by the National Science Foundation’s Emerging Frontiers in Research and Innovation program between MIT CSAIL and the Pliable Math Lab at Harvard.