Soft Robotic Hand With Human-Like Touch

Soft Robotic Hand That Sees Around Corners: FlexiRay and the Future of Human-Like Touch

By Dr. Sagar Rajkuwar, ENT Specialist, Nashik, Maharashtra, India

🌐 www.entspecialistinnashik.com

Introduction

Robots are no longer confined to rigid factory lines performing repetitive tasks. As robotics moves into homes, hospitals, agriculture, logistics, and human-centered environments, the demand for safe, adaptive, and intelligent manipulation has grown dramatically. One of the biggest challenges in this evolution is touch. Humans rely heavily on tactile feedback to grip, feel, adjust, and manipulate objects. Replicating this capability in robots has proven difficult—until recently.

A groundbreaking development from researchers at Zhejiang University introduces FlexiRay, a soft robotic hand that can “see” around corners and achieve a form of human-like touch. Published in Nature Communications (2025), this innovation represents a paradigm shift in robotic sensing: instead of resisting deformation, FlexiRay harnesses deformation to improve perception.

This article explores how FlexiRay works, why it matters, and how it could redefine the future of soft robotics and human–robot interaction.

Why Touch Matters in Robotics

Humans manipulate objects effortlessly because our hands continuously collect information about:

• Pressure and force
• Texture and surface roughness
• Temperature
• Object shape and position
• Finger posture (proprioception)

This rich sensory feedback allows us to pick up a fragile glass, knead dough, or button a shirt without conscious effort. Robots, on the other hand, traditionally rely on vision-based systems, which struggle in situations involving occlusion, poor lighting, or irregular object shapes.

Limitations of Traditional Robotic Tactile Sensors

Most modern robotic hands use visual–tactile sensors that embed small cameras under transparent elastomer skins. While effective in controlled settings, these sensors face key limitations:

• They require stiff materials to maintain camera focus
• Large deformations block or distort the camera’s field of view
• Sensing coverage is limited to small, flat contact areas
• Flexibility is sacrificed for image quality

These constraints make it difficult for robots to handle fragile, soft, or irregularly shaped objects, which are common in real-world environments.

Enter FlexiRay: A Soft Robotic Hand That Sees What It Touches

FlexiRay was designed to solve what researchers call the “blind spot problem” in soft robotic sensors. The key idea is simple yet revolutionary: use the deformation of a soft finger as part of the sensing mechanism rather than treating it as a problem.

Inspiration from Biology

The design of FlexiRay draws inspiration from the human hand and from nature—specifically, the Fin Ray Effect, a structural principle observed in fish fins.

When pressure is applied to a Fin Ray structure, it naturally bends toward the force instead of away from it. This allows the structure to wrap passively around objects, improving grip and contact.

FlexiRay applies this same principle to robotic fingers.

The Fin Ray Effect: Passive Adaptation in Action

The Fin Ray Effect enables FlexiRay’s fingers to:

• Passively conform to object shapes
• Wrap around irregular or fragile items
• Distribute pressure evenly
• Maintain stable contact without complex control algorithms

This passive adaptability is crucial for safe and effective manipulation, especially in environments where robots interact closely with humans.

How FlexiRay “Sees” Around Corners

The most remarkable feature of FlexiRay is its ability to maintain continuous visual perception, even when the finger bends, twists, or wraps around an object.

Multi-Mirror Optical System

Instead of a fixed camera view, FlexiRay uses:

• A single internal camera
• A multi-mirror optical architecture
• Deformation-driven redirection of the optical path

As the soft finger bends inward under pressure, the internal mirrors dynamically redirect the camera’s field of view. This allows the camera to effectively look around corners, maintaining visibility of the contact surface.

In other words, mechanical deformation drives optical perception.

This transforms what was once a limitation—bending and twisting—into a sensing advantage.

Multi-Layered Smart Skin

FlexiRay’s sensing capabilities are enhanced by a multi-layered skin structure, which includes:

• Reflective layers for optical signal capture
• Thermochromic materials that change color with temperature
• Soft elastomers for safe human interaction

This layered design enables the robotic hand to sense multiple physical properties simultaneously.

Five Human-Like Senses in One Soft Finger

Using a combination of mechanical design and deep learning, FlexiRay can perceive:

1. Force – How much pressure is being applied
2. Contact Location – Where the object touches the finger
3. Texture – Surface patterns and roughness
4. Temperature – Heat transfer via thermochromic response
5. Proprioception – The finger’s own shape and deformation

All of this information is extracted from visual data captured by a single camera and processed using deep neural networks.

Deep Learning: Turning Deformation into Intelligence

Raw visual data alone is not enough. FlexiRay relies on deep learning algorithms trained to interpret complex deformation patterns.

The neural networks learn to:

• Map visual changes to force magnitude
• Recognize textures from deformation patterns
• Infer finger posture and contact geometry
• Distinguish between different object materials

This approach allows FlexiRay to achieve over 90% effective sensing coverage, even during large deformations—a major improvement over traditional tactile sensors.

From Avoiding Deformation to Leveraging It

One of the most important contributions of FlexiRay is a conceptual shift in robotics.

Traditional Approach

• Deformation is undesirable
• Sensors must remain flat and rigid
• Flexibility compromises perception

FlexiRay Approach

• Deformation is valuable
• Softness enhances sensing
• Structure and perception work together

This shift opens the door to a new class of intelligent soft robots.

Demonstration: Remote Tea-Making Task

To demonstrate real-world usability, the researchers tested FlexiRay in a remote tea-making task.

The experiment showed that:

• The robot could manipulate delicate objects
• Long sequences of actions were possible
• Multimodal feedback improved teleoperation
• Operators received force, position, and thermal cues

This highlights FlexiRay’s potential not just for autonomy, but also for human-in-the-loop robotic systems.

Key Advantages of FlexiRay

1. Large Effective Sensing Area

Because it conforms to object shapes, FlexiRay senses over a much larger contact surface than rigid sensors.

2. Safety and Compliance

Soft materials reduce the risk of injury or damage during interaction with humans or fragile objects.

3. Simpler Hardware

A single camera replaces multiple embedded sensors, reducing system complexity.

4. Robust Performance

Sensing remains reliable even during extreme bending, twisting, or wrapping motions.

Potential Real-World Applications

1. Healthcare and Assistive Robotics

• Handling medical instruments
• Assisting patients safely
• Rehabilitation and prosthetics

2. Agriculture

• Picking fruits without bruising
• Handling irregular natural products

3. Logistics and Packaging

• Grasping soft or deformable packages
• Sorting objects of varying shapes

4. Domestic Robots

• Household chores
• Cooking and food preparation
• Safe human–robot interaction

Implications for Human–Robot Interaction

FlexiRay’s ability to safely and intelligently interact with the physical world makes it ideal for environments where humans and robots coexist.

Soft, perceptive hands:

• Increase trust in robots
• Reduce accident risk
• Enable more natural collaboration

This is a crucial step toward robots that truly understand their physical surroundings.

Future Directions

The research team plans to:

• Expand FlexiRay into multi-fingered robotic hands
• Integrate imitation learning, allowing robots to learn directly from human demonstrations
• Improve perception accuracy using richer datasets
• Combine tactile data with vision and audio for holistic sensing

By merging advanced hardware with learning-based control, FlexiRay could become a foundation for next-generation dexterous robots.

Why FlexiRay Matters

FlexiRay is more than a new robotic gripper—it represents a new philosophy in robotic design.

By embracing softness, deformation, and biological inspiration, researchers have shown that robots do not need to mimic human rigidity to achieve human-like capability. Instead, intelligence can emerge from compliance.

Conclusion

The FlexiRay soft robotic hand demonstrates that seeing and touching no longer need to be separate senses in robotics. By combining flexible structures, clever optical design, and deep learning, it achieves a level of tactile perception previously thought impossible in soft systems.

As robotics continues to move closer to daily human life, innovations like FlexiRay will play a critical role in making robots safer, smarter, and more intuitive. The ability to see around corners may well be the key to giving robots a true sense of touch.

Reference:
Wang Y. et al., Flexible robotic hand harnesses large deformations for full-coverage human-like multimodal haptic perception, Nature Communications (2025).

Disclaimer

This article is for informational purposes only and does not constitute professional or technical advice.

📞 ENT Consultation & Surgery

Dr. Sagar Rajkuwar (MS-ENT)
Prabha ENT Clinic, Ambad, Nashik
📱 7387590194 | 9892596635
🌐 www.entspecialistinnashik.com