3 ms·

Does someone have a list that is more about actual innovations and less about products not even developed in 2018?

Animats 8y ago
Good point. Universal Robots is doing well. They're in the "collaborative robot" business, like Rethink was before the bankruptcy. Instead of focusing on hype, they've been making good robots that work and shipping them. But the numbers aren't that big - 27,000 total. They still cost too much. There's no good reason a modest sized industrial robot needs to cost more than a car. Other than low volume. Foxconn hasn't been doing as well with their robots as expected. In 2011, they were talking about installing a million robots in their plants over 3 years. They've been installing about 30,000 a year. Outside of automotive, industrial robots are still rather rare. [1] https://www.universal-robots.com/ https://www.universal-robots.com/
- tootie 8y ago
  The e-series are really slick. You're going to see these in a lot of industrial settings and sneaking into public spaces more and more. Price is still high (~$50k for a full kit) but it's very accessible for developers.
Eridrus 8y ago
sim2real research for robotics is super exciting. For all the work Boston Dynamics has done making carefully hand-tuned locomotion controllers, recent work has made that basically learnable in simulation, and then transferable to the real world: https://youtu.be/aTDkYFZFWug https://youtu.be/aTDkYFZFWug
- jcims 8y ago
  That’s super interesting, thank you for sharing. Lex Fridman has a recording of the CEO of Boston Dynamics on his podcast. It was from a lecture Q&A at MIT I believe. And in there he talks about the control challenges as though they are likely the bigger problem to tackle vs. the mechanical (which seems obvious on one hand but interesting to hear him say). In an off-handed comment he said how hooking the mechanics up to a person (so they could run the robot in a fly by wire way) shows just how fast the machine is and how much the software/sensor/controller stacks are slowing them down. He also discusses their simulation capabilities a bit. It definitely sounds like there is some software training going on, but this obviously cranks the fidelity up quite a bit. Would be cool to see how they train it.
- tnecniv 8y ago
  This is neat. Any idea if in the paper they address how well the network works if you swap out the robot for another one of the same model / if the parameters of the physical robot change?
tnecniv 8y ago
First thing that comes to mind for me is practical improvements to sum-of-squares optimization for designing robust controllers, but I'm biased. Really though, I would skim through the awards list and workshop lists of any of the big conferences, e.g. ICRA, RSS, CoRL, WAFR. That will give you some ideas of the big themes lots of people are exploring right now.
- genericone 8y ago
  Good idea, they do often say if Einstein didn't discover relativity, someone else would have within the decade, due to the spread of ideas/inspirations. Taking in the field is likely the best way to discover a 'theme' to which a new innovation will be involved with, however narrow or broad that theme.
  - tnecniv 8y ago
    Yes, also most years in any field don't have a single revolutionary paper to which you can point. The reason I suggest looking at the workshops at conferences is that each one is effectively a cluster of ~6-8 talks and papers / posters with a common theme. Moreover, the ideas for these workshops are submitted, reviewed, and chosen by experts in the field so you know at least someone thinks it's an interesting developing area.

Good point. Universal Robots is doing well. They're in the "collaborative robot" business, like Rethink was before the bankruptcy. Instead of focusing on hype, they've been making good robots that work and shipping them. But the numbers aren't that big - 27,000 total. They still cost too much. There's no good reason a modest sized industrial robot needs to cost more than a car. Other than low volume. Foxconn hasn't been doing as well with their robots as expected. In 2011, they were talking about installing a million robots in their plants over 3 years. They've been installing about 30,000 a year. Outside of automotive, industrial robots are still rather rare. [1] https://www.universal-robots.com/ https://www.universal-robots.com/

tootie 8y ago
The e-series are really slick. You're going to see these in a lot of industrial settings and sneaking into public spaces more and more. Price is still high (~$50k for a full kit) but it's very accessible for developers.

Eridrus 8y ago

sim2real research for robotics is super exciting. For all the work Boston Dynamics has done making carefully hand-tuned locomotion controllers, recent work has made that basically learnable in simulation, and then transferable to the real world: https://youtu.be/aTDkYFZFWug https://youtu.be/aTDkYFZFWug

jcims 8y ago
That’s super interesting, thank you for sharing. Lex Fridman has a recording of the CEO of Boston Dynamics on his podcast. It was from a lecture Q&A at MIT I believe. And in there he talks about the control challenges as though they are likely the bigger problem to tackle vs. the mechanical (which seems obvious on one hand but interesting to hear him say). In an off-handed comment he said how hooking the mechanics up to a person (so they could run the robot in a fly by wire way) shows just how fast the machine is and how much the software/sensor/controller stacks are slowing them down. He also discusses their simulation capabilities a bit. It definitely sounds like there is some software training going on, but this obviously cranks the fidelity up quite a bit. Would be cool to see how they train it.
tnecniv 8y ago
This is neat. Any idea if in the paper they address how well the network works if you swap out the robot for another one of the same model / if the parameters of the physical robot change?

tnecniv 8y ago

First thing that comes to mind for me is practical improvements to sum-of-squares optimization for designing robust controllers, but I'm biased. Really though, I would skim through the awards list and workshop lists of any of the big conferences, e.g. ICRA, RSS, CoRL, WAFR. That will give you some ideas of the big themes lots of people are exploring right now.

genericone 8y ago
Good idea, they do often say if Einstein didn't discover relativity, someone else would have within the decade, due to the spread of ideas/inspirations. Taking in the field is likely the best way to discover a 'theme' to which a new innovation will be involved with, however narrow or broad that theme.
- tnecniv 8y ago
  Yes, also most years in any field don't have a single revolutionary paper to which you can point. The reason I suggest looking at the workshops at conferences is that each one is effectively a cluster of ~6-8 talks and papers / posters with a common theme. Moreover, the ideas for these workshops are submitted, reviewed, and chosen by experts in the field so you know at least someone thinks it's an interesting developing area.