Skip to main content

Useful Robots

· 3 min read
Yuichiro Aoki
Yuichiro Aoki
Developer

High school robotics competitions are popular and have been around for decades. Even a hobbyist can build a robotic arm with a 3D printer. However, why is it so difficult to build a useful robotic arm?

Robitc Arms are Expensive

It’s 2024 and the only useful robots you can buy are still coffee machines and roombas..
Comma AI blog

For the last couple of years, multi-purpose humanoid robots like Atlas from Boston Dynamics has been viral on the internet. Unitree G1 is awesome. It only costs $16,000, but maybe it will take even a decade for it to be useful to the general public.

It's the same reason why mid to small size factories are still using human labor. The cost of a robotic arm is still too high compared to human labor. People here discussed whether solving robotics was a software problem or a hardware problem. Thing is the software is not ready yet and the hardware is too expsensive to replace human labor at this point.

Interesting quote from Comma AI blog:

Humanoid robots are exciting, but incredibly difficult. There are many useful intermediates that must be achieved before useful humanoid robots as a product are realistic, companies that don’t recognize this will fail.

Yes, building useful multipurpose robots is difficult even for big robotics companies with a lot of resources, but starting with one of the specialized robots seems to be a good way to go.

3D Printed Robotic Arms​

As 3D printers become more popular, many people are building robotic arms with 3D printers. That way you can build a robotic arm with a low cost with fast iteration.

However, 3D printed robotic arms are not useful for most cases. They are mostly for educational purposes. Some people sell affordable 3D priinted robotic arms, but you can see the criticism in the comment section of the following video:
Automata aims to "democratise robotics" with $3000 six-axis robot

Aluminium Robotic Arm​

Most of the professional robotic arms are made of aluminium. Aluminium is light and strong, but the cost of the manufacturing process is also high. You need a CNC machine to make the parts. The iteration is slow and expensive.

Of course, it weights more than 3D printed robotic arms, and you need more powerful motors and harmonic drives or planetary gearboxes to move the arm, which makes it heavier and more expensive. With all the cost, the reatail price of a aluminium robotic arm is way beyond the reach of most people.

This is the example of a DIY aluminium robotic arm:

I assume demo in the last part of the video is speeded up by twice. The speed of the robotic arm is not that fast due to the huge reduction ratio of the harmonic drives.

Another example would be AR4. I really hope AR4 to succeed. It's a 6 DOF robotic arm that costs about $2k in total. The max payload is 1kg and the reach is 600mm. It seems they are used mainly for educational and research purposes. I hope they will be useful for the general public in the future.