Jump to content
Join the Unexplained Mysteries community today! It's free and setting up an account only takes a moment.
- Sign In or Create Account -
Still Waters

Boston Dynamics' Handle robot stacks boxes

39 posts in this topic

Recommended Posts

Dark_Grey
Posted (edited)
10 minutes ago, BorizBadinov said:

Its all about available processing power. As micro processors and field arrays become more powerful more possibilities become achievable. In 79 functions were basic and today many times more complex. Those early generation processors had thousands of gates and todays new generations have surpassed 1 billion in a single processor less than 1/4 the size. Also the technology to array logics is common place in multicore processors. 

@Dark_Grey 

Programing a robot to stack a pallet of boxes isn't as complex as picking fruit. A pallet has a repeatable pattern and a set amount of variables.

All in due time :tu:

Picks ripe fruit with 61% accuracy and works 20 hours a day. This is only the first generation. Again I ask, why are we importing all this immigrant labor for jobs that will soon be done by robots exclusively? Politicians only looking at the short term to get votes?

Edited by Dark_Grey
  • Like 1

Share this post


Link to post
Share on other sites
 
BorizBadinov

At this point robots are expensive. But it will come to pass in time that they will be affordable. 

I'm not debating pro or con on immigration. That's a whole can of worms with ulterior motives on all sides that won't be solved here. 

I just like robots. 

61% accuracy means it's just a little better than 50/50. Someone still has to sort through and remove the 39% waste. That's hardly cost effective yet but yes it will get better. Until then someone has to pick the peppers...

Share this post


Link to post
Share on other sites
Dark_Grey
3 hours ago, BorizBadinov said:

At this point robots are expensive. But it will come to pass in time that they will be affordable. 

I'm not debating pro or con on immigration. That's a whole can of worms with ulterior motives on all sides that won't be solved here. 

I just like robots. 

61% accuracy means it's just a little better than 50/50. Someone still has to sort through and remove the 39% waste. That's hardly cost effective yet but yes it will get better. Until then someone has to pick the peppers...

Just for fun, what do you predict for the timeline on near perfect robots? I know it depends on the task and other factors, so for the sake of argument let's concentrate on the fruit pickers. They are at 61% accuracy now but I imagine that is due to the software more than the hardware. Extending an arm with a little saw to cut the branch isn't super complicated as far as robots go. 

I say within 5 years they will be up to 80% accuracy. It's a matter of tweaking the algorithms to better recognize "ripeness". This is facial recognition for produce and facial recognition has improved by leaps and bounds in a very short time.

  • Like 1

Share this post


Link to post
Share on other sites
marsman
18 minutes ago, Dark_Grey said:

Just for fun, what do you predict for the timeline on near perfect robots? I know it depends on the task and other factors, so for the sake of argument let's concentrate on the fruit pickers. They are at 61% accuracy now but I imagine that is due to the software more than the hardware. Extending an arm with a little saw to cut the branch isn't super complicated as far as robots go. 

I say within 5 years they will be up to 80% accuracy. It's a matter of tweaking the algorithms to better recognize "ripeness". This is facial recognition for produce and facial recognition has improved by leaps and bounds in a very short time.

 

 

 

  • Like 2

Share this post


Link to post
Share on other sites
marsman

 

  • Like 1

Share this post


Link to post
Share on other sites
marsman

 

  • Like 1

Share this post


Link to post
Share on other sites
Dark_Grey
15 minutes ago, marsman said:

[video]

So I can give you an (almost) indestructible robot arm with a grip strength of 2500 psi, able to fully rotate the wrist and it can even provide some feedback when touching things but the catch is you have to lose your human arm. What say you?

Share this post


Link to post
Share on other sites
marsman
Just now, Dark_Grey said:

So I can give you an (almost) indestructible robot arm with a grip strength of 2500 psi, able to fully rotate the wrist and it can even provide some feedback when touching things but the catch is you have to lose your human arm. What say you?

 

it can be incorporated to a robot

 

heres the clunky version

 

  • Like 1

Share this post


Link to post
Share on other sites
marsman

The farming robots of tomorrow are here today

 

  • Like 1

Share this post


Link to post
Share on other sites
marsman

 

  • Like 1

Share this post


Link to post
Share on other sites
BorizBadinov
7 hours ago, Dark_Grey said:

Just for fun, what do you predict for the timeline on near perfect robots? I know it depends on the task and other factors, so for the sake of argument let's concentrate on the fruit pickers. They are at 61% accuracy now but I imagine that is due to the software more than the hardware. Extending an arm with a little saw to cut the branch isn't super complicated as far as robots go. 

I say within 5 years they will be up to 80% accuracy. It's a matter of tweaking the algorithms to better recognize "ripeness". This is facial recognition for produce and facial recognition has improved by leaps and bounds in a very short time.

Thanks for the cool links. Don't get me wrong, these robots are all awesome and I love seeing the tech move forward. There is however a time-lag between prototype and production as well as acceptance. 

Things are moving at an incredible pace tech wise, so five years is probably a conservative estimate for 80%. One has to be a little careful about taking these videos and statistics at face value though. Take the Parkour video for instance. Its a little less shiny when you know it took 20 or 30 tries.

You are correct in my opinion, saying software is the hang-up. With robots more trial and error seems required to perfect the systems and care must be taken that it cant go rogue in the case of knife wielding pepper pickers :D It's not going to hold up a convenience store but it might make a mistake involving humans in the work environment. The bigger the sandbox the more variables are introduced.

As more features are added to incorporate more "senses" these tools will soon be able to sense moisture content as well as coloration, weight per volume, location memory,  etc. and likely outperform humans, as well as not caring about work life balance. Hardware has to be created with useable control features that programmers can exploit so software will always lag behind hardware. 

Sometimes simplicity is best as far as the mechanics of the system, but that in turn limits the ability to specific tasks. These task specific robots will apex quickly because they can focus on a single goal and not have the overhead of too much to think about. The ones that mow your lawn and vacuum for example are already quite functional.

True autonomy in multiple environments will require tool use as opposed to just built ins. I think that is a long way off still. Caregiver and rescue robots will likely take another decade to 20 years to be more than novelty in my opinion. In controlled situations bots are beginning to accomplish very cool things but when you add in unknown factors like terrain and weather they still have a ways to go. I have been very impressed though with the balancing and recovery ability of many of the new designs and navigation of obstacles is really the hurdle to surpass. We only get to see the tech that is demonstrated though and not everything is done in public view so I could be way off. I am just a spectator.

Todays AI is programmed to mimic a true thinking entity. When and if that breakthrough happens the game changes dramatically. I believe that will require something more organic.

What I have really been excited about are the breakthroughs in robotic prosthetics. Back in the early 2000s when I read about human cells merged with silicon for the first time, I expected that to progress faster than it has. 

It will be interesting to see the future of robotics for sure.

  

  • Like 2

Share this post


Link to post
Share on other sites
marsman
Posted (edited)
7 hours ago, BorizBadinov said:

Todays AI is programmed to mimic a true thinking entity. When and if that breakthrough happens the game changes dramatically. I believe that will require something more organic.

 

 

look up Sophia.....AI robot....she/it....can have conversations

 

 

 

Edited by marsman
  • Like 1

Share this post


Link to post
Share on other sites
BorizBadinov
13 minutes ago, marsman said:

 

look up Sonia.....AI robot....she/it....can have conversations

 

 

 

I have seen this robot. It's definitely cutting edge automation but far from true ai. This might be as close as we ever get though.

Her responses are from a preprogrammed list some of which were designed for media attention obviously.

  • Like 1

Share this post


Link to post
Share on other sites
BorizBadinov
Posted (edited)

 

Baby X is another interesting AI experiment. 

Edited by BorizBadinov
Link didn't work
  • Like 1

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.