💾 Archived View for sotiris.papatheodorou.xyz › gemlog › 20230308_why_do_we_build_robots.gmi captured on 2023-07-10 at 13:40:36. Gemini links have been rewritten to link to archived content

-=-=-=-=-=-=-

Why do we build robots?

I'm nearing the end of my PhD in robotics and often wonder what the purpose of robots is. But first let me clarify what I mean by robots. A robot is a machine that can autonomously act on and react to its environment. The key word here is “autonomously”. Different kinds of robots have different degrees of autonomy.

Surgical robots are on the lower end of the autonomy spectrum, having practically no autonomy. A surgical robot needs to be continuously operated by a surgeon, like any other tool in the surgeon's disposal. Surgical robots can only act on their environment if commanded to by a human operator. It's fairly easy to find useful applications for human-operated robots, surgery, search and rescue operations in dangerous environments or safe inspection of difficult to access structures.

Industrial robots have a higher amount of autonomy in the sense that once activated they keep performing a task with minimal to no human supervision. However they operate in well-known and highly-controlled environments allowing them to be pre-programmed to perform some very specific tasks. It's not possible to move an industrial robot from one assembly line to another without re-programming it, even if both assembly lines are used for the same kind of product, e.g. different car models. Industrial robots can act on their environment but their reaction to it is typically very limited. This kind of low-autonomy robot is well suited to repetitive tasks that require high accuracy or great strength, which typically means an industrial assembly line. They might still be useful for a society that's not hooked on consumption like ours is.

Robotic vacuum cleaners are a more recent example of a robot with a high degree of autonomy. Similarly to industrial robots they can perform their task without human supervision. However, robotic vacuum cleaners don't require a well-known operating environment. The same robot can vacuum multiple different houses without requiring any modification to its software or hardware. Robotic vacuum cleaners not only act on but also react to their environment by detecting and avoiding obstacles. It's this class of highly autonomous robots that's the topic of my PhD and that I'm struggling to find a purpose for.

The purpose of the highly autonomous robots currently being used and research is to replace humans in certain tasks. Autonomous cars so that Uber doesn't have to pay for drivers, autonomous drones so that Amazon doesn't have to pay for delivery people, autonomous vacuum cleaners so that Heathrow Airport doesn't have to pay for cleaners. Do you see the pattern? Of course capitalism will try to squeeze a profit through any means so it's expected that robots will be used that way too. What do we gain by replacing humans with robots other than improving the profit margins of a handful of companies? Robots that are also bad for the environment since they're full of electronics and batteries.

I think autonomous robots can be worthwhile when they augment humans instead of replacing them. But most of the tasks that could potentially be done by autonomous robots can also be done much more easily by human-operated robots. So far I've only found one case where truly autonomous robots can make sense, environments where human operation is impractical or impossible. What I have in mind here is space exploration where transmission delays can make live teleoperation impractical and underwater exploration where the data rate between a surface and an untethered underwater vessel is severely limited.

Maybe building autonomous robots can be a goal in and of itself, there many interesting problems needing to be solved to get there. I hope there's more to it than that though. Maybe I should just change fields after the PhD.

Sotiris 2023-03-08