The Machine-Robot Continuum

The Machine-Robot Continuum

Reading the article of “What is a robot?” by Matt Simon in Wired [3] got me thinking about the definition of a "Robot".

It took me back to 2010, during my last year of Engineering school. I was in exchange student program in the UK and worked on an innovative project in Augmented Reality (AR). Back then, not many people knew what AR is. People heard about Virtual Reality (VR) but not Augmented Reality. While working on that project I stumbled upon a similar question. What is AR? What is VR? Where one ends and the other begins?

Virtual Reality is defined as a system (VR headsets, projected environments…) that generates realistic images, sounds and other sensations that simulate a user's physical presence in a virtual environment.

Augmented Reality is defined as the contextual integration of computer generated media (text, images, 3D content…) on-top of the real world, directly (like Microsoft’s HoloLens) or indirectly (via a smartphone for example).

Wikipedia (#1 source of information) states that: “ Augmented reality enhances one’s current perception of reality, whereas in contrast, virtual reality replaces the real world with a simulated one. Augmentation techniques are typically performed in real time and in semantic context with environmental elements, such as overlaying supplemental information like scores over a live video feed of a sporting event.” [1]

Things are getting fuzzier at the borders and this is why Milgram and Kishino [1] suggested a continuity rather than a firm definition. They suggested a Reality-Virtuality Continuum. In one side, the 'Real Environment' stands while in the other extreme the pure 'Virtual Environment' stands. Everything between is mixed-reality: augmented reality or augmented virtuality.

Figure 1: Reality-Virtuality Continuum by Milgram and Kishino

But Robots?

Robots seems to rise a similar question between robots and machines. Many roboticists find it hard to come up with a conclusive definition that everyone will agree on.

As Matt Simon wrote “ ask three different roboticists to define a robot and you’ll get three different answers”.

I start to think that maybe a Continuum is more adequate approach. Let me try to explain:

We all agree that the "ultimate" robot is a machine that can move in space, think, perceive, act and communicate. But other less capable machines are also robots like robotic vacuum cleaners, manufacturing robotic arms etc... each has a different set of abilities and a different level of autonomy.

There are many factors of what can make a "robot" a better one: AI, maneuverability, actuations, manipulation, sensing etc. Each is a step towards a more complete system that can think and act in the natural world. I think that we all agree on two things a robot must have: at least some ability to move (otherwise it is just a "device") and some programmable functions. The problem is that some (most?) will say that this definition can be of a simple machine and not enough for a glorious "robot".

Is a dishwasher a "robot"? It has actuators to spray the water, it has different programs that have different steps. It can even sense when it is missing products and communicate it to its owner via dedicated LED.

Must a “robot” be smart? If so, is a simple RC car that can advance by itself and turn when it senses a wall (with simple ultrasonic sensors) smart enough? The first iRobot vacuum cleaners had only bump sensors and they moved randomly in space. Is that smart enough to be called a robot or is it just a machine? What about robotic arms for manufacturing which only move again and again to whatever commands they were programmed to?

This is where I think the continuum comes to play. Assume a straight line in which at one extreme (left side for that matter) you have a basic "machine". Like a dishwasher or a RC car. Very basic capabilities. In the other extreme (the right-hand side) you have a fully articulated artificial intelligence machine (Let’s call it “Physical AI”) that is stronger, smarter and more connected than humans (Terminator-like robot). Points closer to the left-hand side are “advanced machines” or “augmented machines”. Machines that have limited perception and/or actions on the environment. Getting further and further on this line, you will start seeing more capable machines with better sensing capabilities and actuators, better AI and connectivity. Moving from "simple robots" (the first vacuum cleaners for example) to "smart robots" and furthermore to the Physical AI agents which surpass humans in any aspect.

Figure 2: The Machine-Robot Continuum

I think that this approach can simplify the definition. A robot can be anything between a simple RC car or a dishwasher for that matter, to the Terminator (oops, no killer robots! Sorry Elon Musk...).

What’s next

To make it a bit more approachable, I've placed some examples over the Machine-Robot continuum. Some are actual products, some are fictional and some are a category of products (like self-driving cars). As I tried to explain above, those are not firm definitions. I’ve placed these examples based on my own opinion which can be discussed.

It is important to point that I don’t think that robots on the right are ‘better’ than robots on the left. They are generally more autonomous and have more technology in them, but not ‘better’. ‘better’ is very subjective, it is something that has more value to the user. From my own perspective and needs, a dishwasher is much more valuable than an autonomous drone, making it a better product to me.


Figure 3: The Machine-Robot Continuum with examples

Now the question lies, where is your robot on this line?

I will be very happy to hear your thoughts and together maybe we can add more and more examples and perfect that approach of robot definition.


Sources:

[1] https://en.wikipedia.org/wiki/Augmented_reality

[2] P. Milgram and A. F. Kishino (1994). "Taxonomy of Mixed Reality Visual Displays". IEICE Transactions on Information and Systems. pp. 1321–1329. Retrieved 2013-10-17.

[3] “What is a robot?” by Matt Simon. Wired - https://www.wired.com/story/what-is-a-robot

Sources for figure 3:

Amichai Oron

UX/UI SAAS Product Designer & Consultant ?? | Helping SAAS / AI companies and Startups Build Intuitive, Scalable Products.

4 个月

???? ??? ?? ?? ???????? ??? ????? ???? ?????? ???: ?????? ????? ??? ??????? ?????? ??????, ?????? ?????? ??????,?????? ????? ????????. https://chat.whatsapp.com/IyTWnwphyc8AZAcawRTUhR

回复

要查看或添加评论,请登录

Amit Moran的更多文章

  • Monitoring Efficiently Indoor Spaces

    Monitoring Efficiently Indoor Spaces

    A little while ago, I read an article about placing cameras in an indoor space. Now, I like geometrical riddles! Let me…

    2 条评论
  • Intel(R) RealSense(TM) Unity Toolkit

    Intel(R) RealSense(TM) Unity Toolkit

    This is a project I developed together with Nadav Zamir from concept to product. Making it easy to create RealSense…

    1 条评论
  • ???? ????? ????? ????

    ???? ????? ????? ????

    https://www.startisrael.

  • My First Virtual Reality Game :)

    My First Virtual Reality Game :)

    My First Virtual Reality Game :) Check it out: https://play.google.

社区洞察

其他会员也浏览了