By: Yuval Zukerman, Director, Edge AI Partnerships
Sep 24, 2025
Advancing Dexterity in Robotics Through Touch
For decades, robotic manipulation has relied primarily on vision systems powered by camera technology. While effective for many tasks, these systems often fall short when robots need to interact with objects that are small, deformable, or hidden from view. To overcome these limitations, a team led by Dr. Tao Yu and Dr. Wenjie Lu of Analog Devices, Inc. (ADI) is pioneering a new frontier in robotic dexterity: multimodal tactile sensing.
Mimicking the Human Sense of Touch
Humans perform countless tasks each day extensively using touch:
- Grasping soft or fragile items
- Feeling the warmth of objects
- Sensing their texture
- Navigating rooms in the dark
For robots to achieve this level of functionality, they need to feel the world, not just see it. Tactile sensing equips robots with pressure, vibration, temperature, and other sensing modalities that emulate the capabilities of human skin.
“Robots today are largely vision-reliant,” says Dr. Yu. “However, many physical properties—like roughness, softness, or real-time deformation—can’t be perceived by vision alone. Tactile sensing opens new possibilities for robust and adaptive interaction.”
Engineering the Next-Generation Sensor
At the core of this transformation is innovative, industrial-grade sensor fabrication. ADI’s latest tactile sensor prototype has a resolution five times higher than that of a human fingertip. These sensors integrate pressure, vibration, and temperature into a compact, affordable package—enabled by advances in micro-electro-mechanical systems (MEMS) and semiconductor packaging technologies. Additional sensing modalities can be added based on the use case as well.
“We’re leveraging state-of-the-art fabrication processes to build sensors that are scalable and durable,” explains Dr. Lu. “The goal is to create artificial fingers that feel like human ones, and maybe even go beyond.”
Scaling for Real-World Applications
The biggest hurdles in making tactile sensing mainstream are cost, complexity, and sensor measurement consistency. Tactile sensors have traditionally been custom-built and expensive. By applying principles from semiconductor manufacturing, ADI aims to reduce both the cost and size of these components, making them more uniform and viable for widespread deployment.
As AI continues to evolve, these sensors will play a critical role in enabling robots to not only touch but also interpret and react to physical cues in real time, essential for applications ranging from industrial assembly to social robots.
Building Smarter, More Human-Like Robots
With tactile sensing, robots can become more than just tools—they can become intuitive partners in a shared environment. This advancement enables a new class of “physically intelligent” machines capable of nuanced, dynamic interaction with the world around them.
“We’re just at the beginning,” says Lu. “But the pace of development is accelerating and the implications are profound. We envision a future where robots touch, feel, and respond as naturally as we do.”