AndroidMagazine.Com

T1

t1 t1 t1

AI integration to humanoids/androids

Since about one year, the world is excited with the appearance of AI systems that can answer most of our questions like (or even better than) humans. No doubt will we see more improved versions soon, probably including the emergence of AGI – Artificial General Intelligence.

We are also seeing improvements in humanoid robots, (androids) by various firms all over the world, many of which didn’t even exist a few years ago.

And how can these AI systems be integrated into these humanoids / androids?

Let’s try to list main considerations:

Obtaining and processing data from a changing dynamic environment and making real time decisions for am unlimited variety of not predefined tasks will require more powerful sensor, processor and memory (hardware) capabilities.

The real time decisions mentioned above require interface between the AI and robot control system and also integrating AI directly into robot’s control system, which would ultimately controlled by the AI. The AI would also be updated and improved by the experiences obtained by the robot,

The robot will need to have NLP (Natural Language Processing) libraries integrated, to understand and respond to verbal language. This would also entail user interface, as well as cloud based communication.

Terms of use:
Copying or republishing of content of this website is not allowed without written permission from us. Note that we make dated records of our posts using several methods.
The content in this website may contain errors, inaccuracies or be incomplete. Use any content on our website at solely your own risk.
Also see our Terms page for more details.

Figure 01 humanoid

The general purpose humanoid is 5 feet 6 inches, 60 kg and can run 5 hours. Check it out on company website at:

https://www.figure.ai/

Musk’s Optimus can sort cubes, balance on one foot

Elon Musk has just posted a video on his Twitter that shows his Optimus android sorting cubes based on their color even in a dynamically changing setting of cubes. The robot does itself by locating his open limbs accurately in 3D space, according to the video. Optimus can also be seen balancing itself on one foot in an elegant acrobatic move in the video.

What is Zero Moment Point?

Zero Moment Point

Post Date : December 4, 2022

From Newton’s Law, to be at rest or not accelerating, the net sum of external forces acting on a body must be countered with the net sum of reactions in equal magnitude and opposite driection. In 2D, this is true for all three axis, which are the movement in x and y axes and rotation about z axis, which are all perpendicular to each other (x and y axes are the axes of the screen plane and z axis is the 3rd dimension perpendicular to the screen, but the rotation around it still falls on the screen plane so evberything is in 2D). In 3D, we again have x y and z axes but this time we have translation (movement on a line) and rotation about all axes.

To keep a robot stable during its movement, in other words not accelerate (by moving or rotating around a point) the law above therefore applies.

The zero moment point (ZMP) is the point on the robot where the reaction from the ground acts such that, in that given instant, by acting from a particular point at the robots base in a particular magnitude, it cancels out the effects of forces from robots weight plus any inertial forces of the robots mass.

Walking robot for example, will need to move its leg such that, this reaction force can fall on the area of the foot, otherwise the robot will start to fall, since the effects of gravity and inertial forces cannot be countered with an equal and opposing reaction. Needless to say, this becomes a key criteria to achieve for an android’s bipedal walking.

Android robots

An android is a robot that resembles a human form, made with the goal of imitating human actions. Their development goes back a long time. With time, their capabilities are increasing, because of developments in robotics and AI technologies. Below is a list of some of the well known android robots, shown in alphabetical order:

Ameca – Engineered Arts – UK

Asimo – Honda – Japan

Atlas – Boston Dynamics – USA

iCub – Istituto Italiano Di Tecnologia – Italy

Pepper – Softbank Robotics – Japan

Robonaut2 – NASA – USA

Sopiha – Hanson Robotics – Hong Kong, China