Next-generation humanoid ‘Figure 02’ unveiled… “3x faster computational power and complex hand features”

-

Artificial intelligence (AI) robot startup Figure has unveiled its recent humanoid robot, ‘Figure 02’. That is the model that was highlighted in a teaser video last week as “the world’s most advanced robot”.

VentureBeat reported on the sixth (local time) that Figure had unveiled the second-generation humanoid robot ‘Figure 02’ with a more agile and complex design.

In line with this, it has similar physical conditions to the previous model, ‘Figure 01’, but great improvements have been made in each hardware and software.

To begin with, the robot’s outer shell was replaced with a matte black color that accentuates its curves as a substitute of a glossy chrome metal.

The unique 01 model routed and exposed cables outside the chassis to permit for sensor and motor alternative and quick repairs. Nonetheless, Figure 02 routed all cables contained in the legs and arms, protecting them from the encompassing environment.

The battery capability has also been improved by 50% to 2.25 kilowatt hours (KWh), allowing operation for as much as 20 hours. The battery pack has been integrated into the robot’s body, bringing the middle of gravity closer to the robot’s center line. This is claimed to assist improve the robot’s agility.

We also custom-designed and installed a motor to optimize the facility and performance of the joints, allowing it to hold objects weighing as much as 20 kg.

The robot’s hands are especially powerful for handling delicate tasks. Supporting 16 degrees of freedom, it may handle almost the identical tasks as humans. Each finger is operated by an integrated device including sensors and motors.

(Photo = Figure)

Not only the looks, but in addition the brain has been strengthened. It’s explained that it has computational and AI inference performance that’s thrice faster than the previous generation.

This enables robots to perform tasks in the actual world completely autonomously.

It is provided with six RGB cameras and provides an onboard Vision Language Model (VLM) that supports fast visual inference using images and videos acquired from these cameras.

We also created a customized AI model in collaboration with OpenAI, and connected a microphone and speaker to enable conversations between robots and other people as before.

Figure 02 might be deployed on BMW’s production line to perform quite a lot of tasks. BMW is testing ways to integrate the robot into its production line, namely how it can communicate with production line software and human employees.

He’s currently tasked with inserting sheet metal parts into fixtures, a task that requires special dexterity but is described as easy and tiring for humans.

While a release date has not yet been set, the figure hints at expanding beyond warehouses and factory floors.

“Our robots mix human-like dexterity with advanced AI to perform a wide selection of tasks in industrial applications and, within the near future, in homes,” the figure said.

Reporter Park Chan cpark@aitimes.com

ASK ANA

What are your thoughts on this topic?
Let us know in the comments below.

0 0 votes
Article Rating
guest
0 Comments
Oldest
Newest Most Voted
Inline Feedbacks
View all comments

Share this article

Recent posts

0
Would love your thoughts, please comment.x
()
x