BRN Discussion Ongoing

[Jimmy17]

"Conditioned for disappointment" the only real statement to summarise my experience over 5 years and one which shines through beyond all thousands of pages of content on this form!!

 

[Boab]

Apologies if you've already seen/heard this.

https://www.eetimes.com/podcasts/brainchips-ip-for-targeting-ai-applications-at-the-edge/

 

[Bravo]

Figure unveils first-of-its-kind brain for humanoid robots after shunning OpenAI​

Helix introduces a novel approach to upper-body manipulation control.​

Updated: Feb 20, 2025 01:46 PM EST

Kapil Kajal

In a significant move in the AI world, California-based Figure has revealed Helix, a generalist Vision-Language-Action (VLA) model that unifies perception, language understanding, and learned control to overcome multiple longstanding challenges in robotics.
Brett Adcock, founder of Figure, said that Helix is the most significant AI update in the company’s history.
“Helix thinks like a human… and to bring robots into homes, we need a step change in capabilities. Helix can generalize to virtually any household item,” Adcock said in a social media post.


“We’ve been working on this project for over a year, aiming to solve general robotics. Like a human, Helix understands speech, reasons through problems, and can grasp any object – all without needing training or code. In testing, Helix can grab almost any household object,” he added.
The launch of Helix follows Figure’s announcement of its separation from OpenAI in early February.
Adcock stated at that time, “Figure has achieved a significant breakthrough in fully end-to-end robot AI, developed entirely in-house. We are excited to reveal something that no one has ever seen before in a humanoid within the next 30 days.”

A series of the world’s first capabilities​

According to Figure, Helix introduces a novel approach to upper-body manipulation control.
It offers high-rate continuous control of the entire humanoid upper body, which includes the wrists, torso, head, and individual fingers.

This level of control allows for more nuanced movements and interactions. Another important aspect of Helix is its capability for multi-robot collaboration.

It can operate simultaneously on two robots, enabling them to work together on shared, long-term manipulation tasks involving objects they have not encountered before.
This feature significantly broadens the operational scope of robotics in complex environments.
Additionally, robots equipped with Helix can pick up a wide range of small household items, including many they have yet to encounter.

This ability is facilitated through natural language prompts, enhancing the ease of interaction and usability.

Helix also employs a distinctive approach by utilizing a single set of neural network weights to learn various behaviors, such as picking and placing items, using drawers and refrigerators, and enabling cross-robot interaction.

This eliminates the need for task-specific fine-tuning, streamlining the learning process.

Lastly, Helix operates entirely on embedded low-power GPUs, which makes it suitable for commercial deployment. This feature highlights its practicality for real-world applications.

Robots and Helix integration​

According to Figure, current robotic systems struggle to adapt quickly to new tasks, often requiring extensive programming or numerous demonstrations.
To address this, the Figure used the capabilities of Vision Language Models (VLMs) to enable robots to generalize their behaviors on demand and perform tasks through natural language instructions.
The solution presented is Helix, the model designed for controlling the entire humanoid upper body with high dexterity and speed.
Helix comprises System 1 (S1) and System 2 (S2). S2 is a slower, internet-pre-trained VLM that focuses on scene understanding and language comprehension.

At the same time, S1 is a fast visuomotor policy that converts the information from S2 into real-time robot actions. This division allows each system to operate optimally—S2 for thoughtful processing and S1 for quick execution.
“Helix addresses several issues previous robotic approaches faced, including balancing speed and generalization, scalability to manage high-dimensional actions, and architectural simplicity using standard models,” according to Figure.
Additionally, separating S1 and S2 enables independent improvements to each system without reliance on a shared observation or action space.
A dataset of around 500 hours of teleoperated behaviors was collected to train Helix, utilizing an auto-labeling VLM to generate natural language instructions.

The architecture involves a 7B-parameter VLM and an 80M parameter transformer for control, processing visual inputs to enable responsive control based on the latent representations generated by the VLM.

Figure introduces ‘never-seen’ humanoid brain that controls full robot body

Figure launches Helix, an AI that allows robots to recognize language and reasoning like humans and grasp household objects without training.

interestingengineering.com

 

[uiux]

JP2024143096A - Attitude estimation system and attitude estimation method - Google Patents

To estimate a user's posture, including a direction of the user's body, using a small number of sensors.SOLUTION: A posture estimation system comprises a measurement member 1 located at any part of four limbs of a user, and a posture acquisition part 520 for acquiring the posture of the...

patents.google.com

Attitude estimation system and attitude estimation method​

Current Assignee: MegaChips Corp

Abstract​

To estimate a user's posture, including a direction of the user's body, using a small number of sensors.SOLUTION: A posture estimation system comprises a measurement member 1 located at any part of four limbs of a user, and a posture acquisition part 520 for acquiring the posture of the measurement member. The measurement member includes an acceleration sensor 14 and a gyro sensor 15. The posture acquisition part 520 includes a reference coordinate determination part 521 for setting a reference coordinate system of the measurement member based on the user's operation of making the measurement member face a target 3, and an attitude estimation part 522 for estimating an attitude of the measurement member relative to the target by acquiring detection values Da and Dr output from the acceleration sensor and the gyro sensor in response to the user's operation of changing the attitude of the measurement member.

GPT analysis:​

This patent describes a posture estimation system that determines a user's body orientation using a minimal number of sensors. It is primarily designed for gaming, VR, fitness tracking, and motion-based interaction systems.

---

1. Purpose & Use​

The system aims to estimate the posture and orientation of a user’s body efficiently, using a small number of sensors instead of a full-body motion capture setup. This is particularly useful for:

• Gaming – Motion-based gameplay using handheld controllers.
• Virtual Reality (VR) & Augmented Reality (AR) – Enhancing user movement tracking.
• Fitness & Rehabilitation – Monitoring body movement for training or therapy.
• Human-Computer Interaction – Intuitive gesture-based controls.

---

2. Sensor Technologies​

The system uses two key inertial sensors, embedded in a measuring device (such as a handheld controller or a wearable limb sensor):

1. Acceleration Sensor (Accelerometer)
   • Measures movement acceleration in three axes (X, Y, Z).
   • Helps determine tilt and linear motion.
2. Gyro Sensor (Gyroscope)
   • Measures rotational velocity in three axes (yaw, pitch, roll).
   • Tracks rotational movement and orientation changes over time.

These sensors are typically placed in:

• Handheld controllers (left and right hands).
• Wearable devices (e.g., strapped to feet or arms).
• Potential expansion to lower body tracking (e.g., sensors on both hands and feet).

---

3. Processing Technologies & Processor Locations​

The system processes sensor data at multiple levels, using different processors located in the controllers and the game console.

A. Processing at the Controller Level (Embedded Processors)​

Each controller (or wearable sensor) contains an onboard processor that performs initial data collection and preprocessing:

• Location: Inside each controller (or wearable sensor).
• Functions:
  • Collects acceleration and gyroscope data.
  • Filters raw data to reduce noise.
  • Performs preliminary sensor fusion to combine acceleration and rotational data.
  • Communicates with the game console via wireless or wired connection.

B. Processing at the Game Console Level (Central Processing)​

The main computational processing happens inside the game console:

• Location: The game console’s central processor (CPU).
• Functions:
  1. Reference Coordinate System Setup
     • The user performs a calibration motion, aligning the controllers to a fixed target (e.g., display screen).
     • This sets a baseline reference coordinate system.
  2. Posture Estimation
     • The console’s processor integrates accelerometer and gyroscope data from the controllers.
     • Uses sensor fusion algorithms to track movement and correct drift.
  3. Common Coordinate Conversion
     • Since each controller has an independent coordinate system, the console converts them into a unified coordinate system for consistent tracking.
  4. Machine Learning-Based Full Body Estimation
     • The console’s processor runs a machine learning model to estimate full-body posture based on limited sensor data.
     • The model is trained to predict shoulder, arm, and torso positions from hand-held controllers alone.
  5. Adaptive Motion Correction for Different Users
     • The system adjusts for different body sizes by applying acceleration correction algorithms.
     • Example: A child's arm will have different acceleration characteristics than an adult's, so the system scales acceleration values based on user height.

---

4. Advantages Over Traditional Systems​

• Fewer sensors required (no need for full-body tracking suits).
• No waist-mounted sensors needed (orientation is inferred from hand-held devices).
• Cost-effective and power-efficient (less hardware, lower processing demands).
• Machine learning integration allows accurate full-body tracking with limited data.
• Adaptable for different users via automated motion scaling.

---

​

 

[Guzzi62]

FF on the other place found below:

DS IAC JOURNAL 2024 No2

See page 15/16:

A Bioinspired System to
Autonomously Detect Tiny,
Fast-Moving Objects in Infrared
Imagery

https://dsiac.dtic.mil/wp-content/uploads/2024/12/DSIAC_2024_Vol_8_No_2.pdf

The DS IAC journal: The Defense Systems Information Analysis Center (DSIAC) is a component of the U.S. Department of Defense’s (DoD's) Information Analysis Center (IAC) enterprise.

Home - DSIAC

Technical Research and Analysis Services The Defense Systems Information Analysis Center (DSIAC) is a component of the U.S. Department of Defense’s (DoD’s) Information Analysis Center (IAC) enterprise. About Us Technical Focus Areas Advanced Materials Autonomous Systems C4ISR Directed Energy...

dsiac.dtic.mil

 

[DingoBorat]

1X's 3rd iteration of NEO.
I think it's actually more uncanny, the closer it gets to moving like a real person, in a big sock..

These don't use the rigid mechanics of other humanoid robots.