SynTouch LLC has been awarded more than $2.5 million in federal grants for two projects related to tactile sensing.
“Over the past five years, we’ve been exploring application of tactile sensing in prosthetic and robotic hands. We have finally homed in on the killer apps,” said Jeremy Fishel, SynTouch director of research and a speaker at RoboBusiness 2015. “These recent awards will allow us to bring this technology to market.”
“These research grants are highly competitive,” he said. “The two projects cover how touch is used. One is for dexterity, and other is perception.”
Simpler is better for grasping
The U.S. Department of Defense grant is intended “to develop and commercialize our NumaTac sensors to enable intuitive fragile grasping in prosthetic hands,” said Fishel. “Current prosthetic hands are difficult to control, particularly with regards to fragile and deformable objects — everyday objects an amputee needs to interact with such as plastic cups, food, etc.”
RBR50 member SynTouch explored conscious tactile feedback from prosthetic hands but found that it was giving too much information to the wearers. “The BioTac is the only sensor that can sense everything a human can — forces, direction, contact points, vibration, heat flow,” Fishel said.
In a pilot study funded by the National Institutes of Health, SynTouch worked well with amputees in the Bay Area. “We played back temperature, pressure, and vibration, but amputees found that information distracting; they just want to grab [an object],” said Fishel.
What worked was detecting the onset of grip in opposing fingers and automatically adjusting the voltage to the motor similar to how a biological inhibitory reflex works, he explained. “We’re working to commercialize our simpler NumaTac sensors for lower cost and more reflexive actions,” Fishel said.
Most amputees are missing only one arm, Fishel noted, so they’ll use their sound sides for fine manipulation and sensing. They need their prosthetics to be able to grasp or steady objects without thinking about them.
“During two-handed tasks, there are more manipulation challenges,” he said. “To take a cup and pour a bottle of water, they need a good second hand. It just needs to work and to apply sufficient force. The wearer just wants to trust it.”
Frustration and confidence
“One of our frustrations is that many of the mind-controlled neural interfaces and brain-controlled prosthetics in the news are too complex or expensive to be practical,” Fishel told Robotics Business Review. “They raise expectations rather than answer the question of what amputees actually need. The NumaTac sensor is lower cost with volume production.”
“There are lot of people who bought into prosthetics; more people are interested in prosthetic hands than actually need them,” he said. “We need to pay attention to what people need versus what’s cool. In fact, I was once one of those who thought, ‘Let’s reproduce Luke Skywalker’s hand from The Empire Strikes Back, but our philosophy evolved to target what would actually improve lives.”
“Controlling strength takes considerable visual focus,” Fishel noted. “Reflex makes it easy. People can grab things without looking, making it more useful.”
In the biological hand, the amount of muscle force used to grip an object depends on reflexive sensory responses as well as the intentions of the brain. In a myoelectric prosthetic hand, the user controls the motor by changing the amount of activation of remaining muscle in the stump, he said.
“The way our reflex works, when you reduce the gain through control with an inhibitory feedback circuit, the motor force more naturally reflects both the command from the brain and the sensory events in the fingers,” Fishel said. “Instead of [the wearer] staring at the hand to decide exactly when to stop, the hand stops automatically after contact, which turns out to be incredibly intuitive because that’s how hands naturally work when grasping.”
“We did a study over the summer with reference to the human hand,” he recalled. “The subjects had to move 12 eggs or crackers with their sound sides. It took them about 10 seconds. With their current prostheses, it took about one minute. With our prosthesis, it took 12 to 15 seconds — amazing progress. All four subjects were shocked.”
“Another big element is regaining confidence,” Fishel said. “For amputees, it’s often embarrassing not to know whether force will be sufficient or if they’ll crush a fruit when picking it up.”
“This is an important application to many consumer-product developers whose customers make decisions on which products to buy based on how they feel,” said Fishel. “Until now, there have been no machines to quantify this information, and the BioTac is the only sensor in the world that is capable of capturing it.”
SynTouch is developing a machine that can poke and slide over surfaces and measure how it feels in human terms — how rough, cool, etc. Normally, this is an expensive and unreliable process, relying on consumer-preference panels to test things such as skin lotions and fabrics.
“People buy consumer products such as handheld appliances and cars partly based on touch,” Fishel said.
SynTouch plans to develop a standard measurement of machine touch for commercialization, and the company has been working with Fortune 500 companies such as Procter & Gamble Co. on this project.
“We’re working on standardizing on 15 dimensions of touch to characterize surfaces, similar to Pantone for color,” Fishel said. “The field of tribology includes friction and profilometry. The problem is that micrometer measurements don’t correlate with human perception. Mechanical probes should have similar properties to a finger to replicate human impressions.”
Los Angeles-based SynTouch wants to measure multiple dimensions simultaneously.
“We want to measure everything a person can feel, and after testing 600 surfaces, we have yet to encounter a surface that people can distinguish but we can’t,” said Fishel.
The BioTac sensors were designed eight years ago, but its sensors have recently improved, and applications are evolving, according to Fishel.
Although most of SynTouch’s customers for the past four years have been academic and industrial researchers, there are definite commercial applications for improved machine sensing.
More on Prosthetics and Robotic Grippers:
In addition to product development, standard tactile measurements could also be good for quality control, fruit picking, or packing boxes in logistics facilities, Fishel said.
“We’re scaling up for the commercial market,” said Fishel. “We’re proud of the hard work of our team on our grant research and on the BioTac products that we make and sell. Up to this point, we’ve managed to avoid needing outside investors, but that could change.”