Figure 2 Humanoid Robot unveiled with new design and capabilities

AI robotics company Figure has unveiled its second generation humanoid robot – Figure 2 with a clean design, smooth autonomy, and numerous new features.

The first impression of the design gets full marks as it advances beyond the first version. It has redesigned joints and parts, which makes it more futuristic of what a humanoid robot should look like.

From foot to knees and torso, all of these parts are designed to work in synchronization and the level of detail to cover the internal architecture is entirely new.

The torso is revamped with new hinges to support a higher degree of movement than the initial design. The figure says that the wiring has been integrated for power and compute communication for concealed wires, higher reliability, and higher packaging.

The biggest upgrade in this new design is the 4th-generation hands. Figure 2 robot comes with human-like fingers and smoother movements with 16 degrees of freedom.

It enables the bot to bend fingers and grab with “human-equivalent strength” and makes it useful for a range of different tasks. Prior to this, the first-gen model didn’t have such a degree of movement and the actions were more robot-like.

Figure 2 Robot 4th Gen Hands (source – Figure)

The head consists of 6 cameras using AI algorithms to detect and act based on input. These help the bot to understand the objects with RGB color array.

The company has upgraded the onboard communicating system and added 3 times the computation and AI inference compared to previous technology. This upgrade allows the bot to react fast and remain efficient.

Figure 2 Robot Head with 6 RGB Cameras (Source – Figure)

The second version utilizes the power of AI models for both vision and speech recognition. In terms of speech, both have conversation capabilities to understand a human with onboard microphones and reply with speakers.

These features come directly from the model trained by the AI company – OpenAI. On the other hand, the vision language model helps the system to comprehend objects and learn new moves with visual information.

Figure 2 Robot VLM and Speech-to-Speech model (Source – Figure)

What’s more important is that the new Figure 2 is unveiled with a 2.25kWh custom battery pack. It is installed in the torso section and provides more than 50 percent energy to improve the runtime.

(source – Figure)