English

Technology

See, touch, and determine the force. Introducing the three technologies that support this loop.

Technology

See, touch, and determine the force. Introducing the three technologies that support this loop.

Technology

See, touch, and determine the force. Introducing the three technologies that support this loop.

TECHNOLOGY

Three technologies behind the grasping hand.

Touch, which reads contact at the fingertips; force control, which decides the force from what it reads; and learning, which gets more accurate the more it’s used on site. Together, they let our hands handle objects of any shape or hardness—without dropping or crushing them.

01

Sense the slip before it happens.

Pressure sensors arrayed across a 3 mm square on the fingertip detect the uneven pressure that appears just before an object starts to slip—and turn it into a cue to re-grip before it drops.

368 points / 1 kHz

Sensing points / Sampling rate

02

Move by force, not by position.

Instead of “how far to move,” the target is “how much force to apply.” Soft objects stay intact; hard ones are held securely.

1ms/8ms

Command cycle / Response latency (avg.)

03

The more it works, the better it fits your site.

The model is updated gradually with contact data gathered during operation, bringing its grip closer to what suits the objects handled on that site.

2–4 weeks after deployment

Typical time to adapt to a site

HOW IT WORKS

One revolution,
in 1 millisecond.

The camera captures the position, the fingertips read the contact, the controller determines the force, and the hand moves. If these four components are separated into different devices, delays occur at the connection points. AWAI's hand runs all four on the same board.

01

tactile sensation

1 / 3

Notice the slip before you slide.

We have arranged 368 pressure-sensing elements within a 3mm square area at the fingertip. Just before an object begins to slip, the pressure on the contact surface shifts slightly. By reading this shift at 1kHz, we convert it into a signal to regrip the object before it drops.

Please enter a description of the image in the alt attribute. This is a sample image.

1.1 Detection Mechanism

Sensing elements 0.15 mm apart feel pressure shift toward the center as a slip begins.Sensors spaced 0.15 mm apart detect the shift in pressure toward the center as slipping begins.

Milliseconds before the object moves, the grip tightens just enough—so soft items keep their shape.

01

tactile sensation

1 / 3

Notice the slip before you slide.

We have arranged 368 pressure-sensing elements within a 3mm square area at the fingertip. Just before an object begins to slip, the pressure on the contact surface shifts slightly. By reading this shift at 1kHz, we convert it into a signal to regrip the object before it drops.

A person's hand touching snow. A photograph capturing the sensation of touching something soft and malleable.

1.1 Detection Mechanism

Sensing elements 0.15 mm apart feel pressure shift toward the center as a slip begins.Sensors spaced 0.15 mm apart detect the shift in pressure toward the center as slipping begins.

Milliseconds before the object moves, the grip tightens just enough—so soft items keep their shape.

01

tactile sensation

1 / 3

Notice the slip before you slide.

We have arranged 368 pressure-sensing elements within a 3mm square area at the fingertip. Just before an object begins to slip, the pressure on the contact surface shifts slightly. By reading this shift at 1kHz, we convert it into a signal to regrip the object before it drops.

A person's hand touching snow. A photograph expressing the sensation of touching something soft and mutable.

1.1 Detection Mechanism

Sensing elements 0.15 mm apart feel pressure shift toward the center as a slip begins.Sensors spaced 0.15 mm apart detect the shift in pressure toward the center as slipping begins.

Milliseconds before the object moves, the grip tightens just enough—so soft items keep their shape.

02

Power Control

2 / 3

Move by force, not by position.

Many robots target "how far to move." AWAI's hand targets "how much force to push with." It doesn't crush soft objects and securely holds hard ones. It determines the force on the spot, adapting to each object.

Please enter a sentence describing the image in the alt text. This is a sample image.

2.1 Control concept

Force-based control holds objects with the intended force, even if size or position is slightly off.

Each joint gets a command every 1 ms, with an 8 ms average delay from sensor to motor.

02

Power Control

2 / 3

Move by force, not by position.

Many robots target "how far to move." AWAI's hand targets "how much force to push with." It doesn't crush soft objects and securely holds hard ones. It determines the force on the spot, adapting to each object.

A photograph of a robot hand gently pinching the stem of a single red gerbera.

2.1 Control concept

Force-based control holds objects with the intended force, even if size or position is slightly off.

Each joint gets a command every 1 ms, with an 8 ms average delay from sensor to motor.

02

Power Control

2 / 3

Move by force, not by position.

Many robots target "how far to move." AWAI's hand targets "how much force to push with." It doesn't crush soft objects and securely holds hard ones. It determines the force on the spot, adapting to each object.

Please enter a description of the image in the alt attribute. This is a sample image.

2.1 Control concept

Force-based control holds objects with the intended force, even if size or position is slightly off.

Each joint gets a command every 1 ms, with an 8 ms average delay from sensor to motor.

03

On-the-job learning

3 / 3

The more you use it, the more it becomes an extension of your hand on-site.

Even for the same task, the items handled and their placements vary from site to site. By gradually updating the model using contact data collected during operation, the system adjusts to the optimal grip force suited to the specific items at that location within a few weeks of installation.

Please enter a sentence describing the image in the alt text. This is a sample image.

3.1 Learning Flow

Every grip, including re-grips, becomes training data, and the model is updated overnight.

After a safety check, updates go live gradually; within 2–4 weeks, re-grips become rare.

03

On-the-job learning

3 / 3

The more you use it, the more it becomes an extension of your hand on-site.

Even for the same task, the items handled and their placements vary from site to site. By gradually updating the model using contact data collected during operation, the system adjusts to the optimal grip force suited to the specific items at that location within a few weeks of installation.

A photograph of an engineer operating a laptop in front of testing equipment lined with wiring and speakers.

3.1 Learning Flow

Every grip, including re-grips, becomes training data, and the model is updated overnight.

After a safety check, updates go live gradually; within 2–4 weeks, re-grips become rare.

03

On-the-job learning

3 / 3

The more you use it, the more it becomes an extension of your hand on-site.

Even for the same task, the items handled and their placements vary from site to site. By gradually updating the model using contact data collected during operation, the system adjusts to the optimal grip force suited to the specific items at that location within a few weeks of installation.

Please enter a sentence describing the image in the alt text. This is a sample image.

3.1 Learning Flow

Every grip, including re-grips, becomes training data, and the model is updated overnight.

After a safety check, updates go live gradually; within 2–4 weeks, re-grips become rare.

SPECIFICATIONS

Main Specifications

Item

Value

Conditions

pressure-sensitive element

368 points / finger

Per 3mm square

Reading cycle

1kHz

All elements simultaneously

Detectable pressure difference

0.8 kPa

Room temperature 25 °C

Item

Value

Conditions

pressure-sensitive element

368 points / finger

Per 3mm square

Reading cycle

1kHz

All elements simultaneously

Detectable pressure difference

0.8 kPa

室温25℃

Item

Value

Conditions

pressure-sensitive element

368 points / finger

Per 3mm square

Reading cycle

1kHz

All elements simultaneously

Detectable pressure difference

0.8 kPa

Room temperature 25 °C

PUBLICATIONS

Papers and Patents

2026

Stabilizing Gripping via Fast Detection of Slip Signs / International Conference on Robotics

2025

Manufacturing method of fingertip-distributed pressure sensor / Patent application

2025

On-Site Adaptation of Grasping Force Model Using Tactile Data / Domestic Conference

2026

Stabilizing Gripping via Fast Detection of Slip Signs / International Conference on Robotics

2025

Manufacturing method of fingertip-distributed pressure sensor / Patent application

2025

On-Site Adaptation of Grasping Force Model Using Tactile Data / Domestic Conference

2026

Stabilizing Gripping via Fast Detection of Slip Signs / International Conference on Robotics

2025

Manufacturing method of fingertip-distributed pressure sensor / Patent application

2025

On-Site Adaptation of Grasping Force Model Using Tactile Data / Domestic Conference

Please include a description of the image in the alt text. This is a sample image.
Please enter a description of the image in the alt attribute. This is a sample image.
A photo of two people looking into a laptop screen and talking with a smile.

CONTACT

Why not try it
at your own site?

If you tell us about the items you handle and your operations, we can prepare a demonstration with actual equipment. You can also send us your own parts or products so we can verify in advance whether they can actually be gripped. Feel free to contact us even before making a decision to implement, or just with technical questions.

Create a free website with Framer, the website builder loved by startups, designers and agencies.