3

A friend and colleague of mine who studies robotics says that bipedal robots present much greater challenges for stability and locomotion than those with more legs.

So why is there so much effort to develop bipedal robots? Are there any practical advantages?

Of course, I see the advantage of having arm-like free appendages, but it seems like 4 legs and 2 arms would generally be a better design.

Ben
  • 5,855
  • 3
  • 28
  • 48
capybaralet
  • 131
  • 4
  • A lot of responses are saying that environments are designed for humans, and so humanoids are the best solution.

    It's a valid argument, but, IMO, not that convincing without more support. It could just point to a lack of imagination.

    – capybaralet Jan 04 '15 at 15:50
  • I like the "science" argument: that studying the problem teaches us useful things. – capybaralet Jan 04 '15 at 15:51
  • Please do not edit your question this way. It changes the context, and makes the answers less applicable. If you knew that to begin with, why ask the question? – Spiked3 Jan 05 '15 at 17:39
  • Well, my question is basically: "Why only two legs?"

    I didn't think people would be silly enough to think that I meant "Why arms?"

    – capybaralet Jan 08 '15 at 15:06
  • Honest question: What alternative would I have? Ask another question (I assume it would be flagged as a duplicate)? Just use comments (then less people read it and I get more answers I consider irrelevant). – capybaralet Jan 08 '15 at 15:08

5 Answers5

3

I would say:

  1. Compactness of the structure: imagine humanoid robots that are required to operate in household environment: two-legged robots might have less difficulties in navigating among tables and chairs standing very close each other.

  2. Energy consumption: of course the power needed increases proportionally with the number of legs.

  3. Human likeness: an aspect that we don't have to neglect if one day we want to have robots among us, both for social reasons and technological reasons. Regarding the latter, generally speaking, copying the human body entails a number of benefits for robots that are designed to act in human context: this applies equivalently to the robot upper-body (e.g. manipulation) and to its lower-body (e.g. locomotion).

Ugo Pattacini
  • 4,005
  • 1
  • 14
  • 36
  • I'm struggling to think of an environment that humans can navigate and dogs cannot.

  • Really? You could of course use smaller legs if you had more of them.

  • – capybaralet Jan 03 '15 at 21:03
  • Simply think of a dog-shaped robot that needs to navigate and at the same time grab objects from the top of the table: the robot will be definitely bigger. The task shapes the body. Besides, a smaller dog, to save energy, has not point in that. – Ugo Pattacini Jan 03 '15 at 23:02
  • However, just as conclusive remark: stabilization and locomotion of bipedal robots is no longer a big challenge to tackle... – Ugo Pattacini Jan 03 '15 at 23:07
  • 1
    @user2429920 The DARPA challenge specifically includes driving/operating a car/vehicle, a perfect example environment where 4 legs would be cumbersome. Not saying cant be done, but a good example where fitting in helps. – Spiked3 Jan 09 '15 at 23:38
  • @Spiked3 - OK, gtk, but that just pushes the question on to DARPA... – capybaralet Jan 11 '15 at 00:44
  • @user2429920, I think you overestimate the difficulty of design of 2-legged compared with 4-legged robots (static walking for humanoid is a solved problem in literature) and you underestimate energy issues plus the mechanical burden of two more legs. Aside from challenges in science, also engineering does matter. You said lack of imagination? There is used to be plenty of imagination behind androids in human history (this somehow pushes the research) while nowadays there is too much imagination around robotics without solid background. – Ugo Pattacini Jan 11 '15 at 10:04
  • Two-legged lower-bodies represent the minimum structure required to accomplish the task and they work quite well; this shows that we are not so bad at solving the dynamics underneath. So, why should we prefer 4 legs (unless the robot needs to carry very very heavy loads)? You do start from the questionable opinion of a friend. This way we could even end up discussing wheels vs. legs... – Ugo Pattacini Jan 11 '15 at 10:10
  • Well, there's no reason for me to challenge his opinion, since he studies it, and I don't. I'm still not convinced that more legs means more energy or 'mechanical burden' (what does that mean, actually?) Like I said, you could use more smaller legs, right? What am I missing here? – capybaralet Jan 12 '15 at 04:07
  • As I tried to explain before, it's the task that usually shapes size the most; therefore, if the task of the robot is to navigate and pick up objects from a table you cannot reduce the size of the legs to save energy, since you then fail meeting the pick-up requirement because the robot is not tall enough. Btw, take a look here: more than 1 year ago, nice 2-legs stabilization, isn't it? – Ugo Pattacini Jan 12 '15 at 18:03