A recent breakthrough in robotics could help amputees regain some of their lost capabilities. The new prosthetic hand design combines layers of sensors with a hybrid robotic structure and machine learning algorithm that reads neuromorphically encoded signals to achieve human-like capabilities. Here’s what you need to know.
The Science of Gripping
When you reach down and pick something up, it may seem like it happens automatically. The reality is that this simple action, which seems like nothing, is a complex interaction between millions of skin mechanoreceptors, your soft tissue, bones, joints, and brain.
Your hand has four primary tactile mechanoreceptors called Merkel cells, Meissner corpuscles, Ruffini endings, and Pacinian corpuscles. The outer layer of your skin has Merkel cells that are designed to respond to light touch.
Next, the Meissner corpuscles sense low frequencies. Followed by the Ruffini endings that determine deformation on surfaces. The final layer, Pacinian corpuscles, detects pressure and high-frequency vibrations.
It’s this mixture of sensors, bones, and tissue that allows humans to easily and quickly sense diverse and complex surfaces. Notably, this structure allows you to pick up an egg without breaking it or to notice that the paper cup of hot cocoa you ordered from the vending machine is starting to slip.
Issues Replicating Human Grips
There have been many attempts to create robotic hands with the same versatility as human limbs. However, all have fallen short to date for several reasons. However, the attempts that have achieved the best results have utilized soft robotics.
Soft Robots differ from traditional robotics in that they lack a hard structure. These devices have seen increased usage across a variety of tasks, including disaster relief and mineral exploration, where their non-conforming design allows them to alter shape to fit through tight spaces. Sadly, it’s this lack of rigidity that has also caused soft robots to fall short of replicating human limbs.
For one, every sensor you add to a flexible soft robot impairs its primary capability, flexibility. Additionally, the majority of these systems are only capable of sensing that touch has occurred. This approach is a far cry from the massive amount of sensory data your brain processes every time you touch an item.
Details of the Study
Recognizing these limitations, a group of scientists from Florida Atlantic University and other leading institutions published the study “A natural biomimetic prosthetic hand with neuromorphic tactile sensing for precise and compliant grasping“1 in Science Advances. The paper looks into a novel hybrid prosthetic hand design that leverages soft robotic joints, a rigid endoskeleton, and a multilayered sensor system.
Hand Design
The engineers created a new hand design that replicates human hands. The multi-finger system uses rubber-like polymers for the creation of the fingers and opposable thumb. The design leverages soft, air-filled finger joints that are actuated with the forearm’s muscles via electromyography. This approach allows the wearer to control it like their real hand.
Source – Science Advances
Hybrid Biomimetic Finger
The hybrid biometric finger design integrates three independently actuated soft robotic joints. These joints were made from Dragon Skin 10 silicone to simulate human skin. The fingers feature 14 independently actuating joints split like your hand, with 3 in each main finger and two in the thumb.
The scientist chose to utilize pneumatic networks to actuate the hybrid finger joints. Specifically, the actuators pressurize air, which inflates them, causing them to move accordingly. This strategy eliminates the need for additional motors and actuators, reducing costs and weight.
The human-like configuration of the soft actuators with the rigid endoskeleton provides more force than a traditional soft robotic option. As such, they determined that the pressurized actuators provide direct force transmission to the manipulated object in a way that allows it to provide precise pressure to specific areas to manipulate the object without damage.
Prosthetic Hand Fingertip
The soft silicone fingertip has an array of multi-layered tactile sensors that allow it to sense pressure and high frequency. It’s the most sensitive part of the hand and is connected to other sensors to provide an in-depth interpretation of the environment and object to be manipulated. Like you, the robot can run its fingertip across a surface to determine a lot of characteristics about an item’s makeup and how it must be handled.
Skeleton
At the core of this approach is the belief that a hybrid design incorporates a rigid 3D-printed rigid endoskeletal structure. The skeleton used by engineers was developed utilizing polylactic acid (PLA). It offers stability, force multiplication, and support for the core components of the hand.
Sensors
The scientists found that they only needed to replicate three layers of tactile sensors to achieve near humanlike performance. Their layout provides reliable tactile feedback in real time to the system, allowing it to make complex decisions to determine the composition, force needed, and approach to successfully manipulate items.
Outer Layer: The outer sensor layer was designed to operate like your epidermis. Your skin can notice the slightest touch. To accomplish this task, the team integrated an outer piezoresistive sensing layer on the surface of the hybrid fingertip. This layer integrates nine tactile sensors. Each sensor only has a total size of 4 mm2 and is spaced 2.5 mm apart, providing full coverage of the fingertip.
Middle Layer: The middle sensing area was built to operate like the Ruffini endings in your body. To accomplish this task, the team embedded a piezoresistive sensing layer. Notably, piezoresistive sensors change electrical resistance when an external force is applied to them.
Interestingly, this layer has six sensors that measure 6 mm2. The team spaced these sensors 2.5 mm apart and placed them in an offset manner from the outer sensor layer to improve tactical reception capabilities.
Inner Layer: The engineers designed the inner layer of the hand to operate like Pacinian corpuscles, detecting high-frequency vibrations and transient pressure from the environment. It was constructed using a 10-mm piezoelectric transducer attached to the fingernail. Specifically, it sits between the soft and rigid components of the hybrid fingertip.
Whenever forces are detected, it generates a small voltage that allows the system to make adjustments and determine the best way to manipulate the object. Additionally, it utilized the rigid fingernail to pick up vibrations in surfaces.
Machine Learning Algorithms
All of this data gets fed into an ML algorithm that gathers, processes, and neuromorphically encodes the data before sending it back to the robotic appendage. The system can utilize neuromorphic responses in relation to its proprietary machine learning algorithm to classify texture.
Interestingly, the system neuromorphically encodes data relative to the mechanoreceptors in human skin using the Izhikevich neuron model framework. This strategy allows the unit to provide naturalistic tactile sensory information through nerve stimulation, which is a first for hybrid robotics.
The Robot Hand Knows What It’s Touching
This strategy enables the robotic arm to determine what it’s touching. The signals bridge the brain and nerves, allowing a wearer to distinguish objects of various shapes and surface textures with ease.
Prosthetic Hand Testing
Testing of the prosthetic hand began with individual fingers. Each finger was tested, and each sensory layer underwent evaluation. Once the team determined that all devices operated independently as planned, they were combined, and the next phase of testing began.
As part of this testing phase of the research, the engineers attached the hand to the UR5 robot arm. From there, the team began attempting to manipulate items. In total, 15 everyday items were selected. The items tested ranged from pineapples, metal water bottles, and soft toys, all the way to a paper cup full of water.
Results and Observations
The test results showed real promise for this technology. The testing provided some insight into its capabilities. In terms of flexibility, the hybrid biomimetic finger achieved 127° of curvature and a 230° angle of flexion with no points of failure.
Additionally, the robotic arm showed versatility and was able to adjust its grip on the fly. In one instance, it only used 3 fingers to grasp a paper cup to avoid bending it and spilling the water. Impressively, the robotic arm’s sensors classified items based on touch with 98.38% accuracy. This rate surpassed both soft robotic and rigid prosthetic fingers, delivering human-like accuracy.
Benefits of the Findings
The benefits that this study brings to the market could change robotics for the better. The hybrid technology demonstrated here could help to improve safety in environments where robots and employees work side by side. Imagine bumping into your robot coworker and them pulling away and apologizing.
Increased Dexterity
The upgraded prosthetic arm showed high dexterity compared to predecessors. It could accomplish tasks that both soft and rigid robots couldn’t. In one example, it was tasked with grasping a ball by conforming around it. It accomplished this task, which would be impossible for a rigid prosthetics.
More Natural
Another major benefit of this style of prosthetic is that it feels more natural to the patient. People suffering from upper limb loss can feel as if they can’t reintegrate into their normal tasks because of a fear of their prosthetic causing harm or injury. This tech promises to allow them to safely interact with their loved ones without concern of hurting them.
Prosthetic Hand Study Researchers
The prosthetic hand study was led by Wen-Yu Cheng of Florida Atlantic University. Other researchers who contributed to the project include Jinghua Zhang, Ariel Slepyan, Mark M. Iskarous, Rebecca J. Greene, Rene DeBrabander, Junjun Chen, and Arnav Gupta of the University of Illinois Chicago.
Interestingly, this same team was the first to introduce electronic skin to the robotics sector in 2018. Now, they have built on this technology further to create capable prosthetics that provide human performance. Their plans include furthering their systems by integrating more sensors, better materials, and increased gripping force.
Real-World Applications & Timeline for Prosthetic Hand Tech
This advancement holds significant promise for individuals with upper-limb loss, offering them the ability to interact with their environment more naturally and safely. Future prosthetics could integrate this technology and provide life-like responses. The same technology could also help improve surgical robots as well.
While currently in the research phase, such prosthetic technologies could become commercially available within the next 5 to 10 years, depending on further development and regulatory approvals. Here are a few other real-world applications for this soft robotic gip tech.
Industrial
The industrial sector has seen a strong shift towards robotics over the last 5 years. This latest technology could help to drive adoption further. Manufacturers are constantly looking at ways to integrate robotics with human workers to improve efficiency without losing quality.
Hybrid robotic features like the one discussed in this study could work alongside humans with less risk. They could also accomplish traditionally human-only tasks like sorting delicate fruit or products like glassware without causing damage.
Agricultural
Farming is another area where robots have found a home. These devices could help to improve harvest by helping to monitor the crop’s health via sensors and ensuring the ripe crops get picked in a timely manner. In the future, soft robots could handle much of the farming process, from planting to picking to sorting good and bad crops.
An Innovative Company Leading the Robotics Industry
The robotics sector continues to boom as more firms and technologies enter it. The future of robotics is bright, and several firms have secured strong positioning in the market. These companies have poured billions into the research and development of more agile, capable, and longer-lasting robots. Here’s one company pioneering robotics.
Ekso Bionics Holdings, Inc. (NASDAQ: EKSO)
Ekso Bionics Holdings Inc. (EKSO +5.66%) entered the market in 2005, seeking to enhance the field of exoskeleton technology and robotic rehabilitation devices. Since its launch, the company has secured numerous high-level contracts focused on further developing its exoskeleton products.
Exoskeletons are robots that are worn by humans. Their design is meant to supplement and enhance your motions. As such, they could be used in factories to help prevent fatigue or on a battlefield to give soldiers more carrying capacity.
Ekso Bionics Holdings, Inc. (EKSO +5.66%)
Ekso Holdings offers several innovative solutions designed to improve the quality of life of patients suffering from the loss of limbs. These products, alongside its positioning and market history, make Ekso Bionics Holdings a smart addition to your portfolio.
Latest on Ekso Bionics Holdings
Prosthetic Hand Study and the Future of Hybrid Robotics
This study demonstrates how nature may have already found the best solution for many design problems. As more engineers seek to mimic nature, their robotic designs will usher in a new age of efficiency. This tech could help ensure that robots can work alongside humans safely and provide additional services that improve both the workers’ lives and the products delivered.
Learn about other cool Robotics Breakthroughs Here.
Studies Referenced:
1. Sankar, S., Cheng, W.-Y., Zhang, J., Slepyan, A., Iskarous, M. M., Greene, R. J., DeBrabander, R., Chen, J., Gupta, A., & Thakor, N. V. (2025). A natural biomimetic prosthetic hand with neuromorphic tactile sensing for precise and compliant grasping. Science Advances, 11(10), eadr9300. https://doi.org/10.1126/sciadv.adr9300