arm slaved robots

=robots =control



Robots with arms that mimic the arm movement of a human operator have been a popular idea for many decades now.

They've been in movies, for example the power loader in Aliens.

Prototypes have been built, for example GE's Handyman robot.


But they aren't used in industry now. Why?


A1) That control system is more expensive than joysticks.

A2) Your arm gets tired if you hold it out for a while, and that makes your accuracy worse. This "gorilla arm" effect is one reason why big touchscreens and gesture controlled computers have been mostly relegated to science fiction.

A3) Robotic arms are generally less flexible and slower than human arms, and this can be frustrating.



What are some potential advantages of robots being controlled by slaving their movement to the arm of a human operator, compared to controlling a hydraulic excavator arm with joysticks?


B1) it's more intuitive

B2) better accuracy at high movement speeds

B3) force feedback

B4) 1 person can control 2 arms better


(B1) doesn't matter for things people use a lot for work. Hydraulic excavator controls are more intuitive than keyboards or manual car transmissions, and people manage to use those quite well. Humans are adaptable.


(B2) doesn't matter for hydraulic excavators because they don't move very fast.


(B3) doesnt matter for hydraulic excavators because their movement is controlled by valves that throttle hydraulic fluid flow, meaning that position is controlled but applied force is not.


(B4) may be relevant, but what's the application you need 2 arms for instead of a single bigger arm? The only current reasons to have one person controlling 2 large arms are very niche.



What might change some of these factors?


(A1) Cheaper microcontrollers and sensors have made it cheaper to monitor an operator's arm movements.


(A2) This means that exactly mimicking arm movement is never going to be practical, but matching arm and hand movement can still be part of a robotic arm control system that might be developed at some point.


(B2) (B3) Different drive systems, such as (electrohydraulic drives, electric motors with strain wave gears, or just a different type of hydraulic design) might be faster or allow force to be controlled as well as position.


(B4) Factories often have robotic arms that repeat the same task, which don't need to be controlled. Factories often have overhead cranes, which sometimes follow as a person guides the cargo horizontally. Those do their job better than a robotic arm would. Hydraulic excavators work well, and a different control system wouldn't improve their performance. The existing tools that factories and mines use handle current applications fairly well, which is why the % of people working in manufacturing and mining is now fairly low.


For a robot mimicking human arm movement to be useful, it would need to be doing a task that's too complex to control it with joysticks, not repetitive enough to completely automate its movement, and valuable enough to be worth paying an employee for.


Do such tasks actually exist? Maybe. For low volume manufacturing work, such as making aircraft, there might be some reasons to use that type of control system for robotic arms.





 back to index