Software is the fin...
 

Software is the final barrier.  

  RSS

casey
(@casey)
Reputable Member
Joined: 8 months ago
Posts: 328
2020-01-19 11:55 am  

When it come to robotics the electronics and its associated hardware required is mainly a question of money. It is already good enough for some amazing robots if only they had a brain.

"... a PR2 robot can clean a full room when remote controlled by a human. The mechanics to do that is already on place and working. The electronics too. The only point missing is the software, and that is why some person must remote control it."

On not using ROS for your robotics startup

 


Quote
frogandtoad
(@frogandtoad)
Reputable Member
Joined: 8 months ago
Posts: 365
2020-01-19 12:10 pm  

@casey

Hmm, that video was uploaded in 2011... I still don't see any robots of today with such capability, even if manually controlled?


ReplyQuote
Robo Pi
(@robo-pi)
Noble Member
Joined: 9 months ago
Posts: 1462
2020-01-19 12:53 pm  

I agree with Mr. Tellez.     But he's talking about how a startup business should be approached.   He specifically addresses the difference between a business and a hobby:

Mr. Tellez, "Again, are we talking about building a business or a hobby? If it is a business, you cannot build your product from scratch, develop all the fancy electronic boards and mechanics, and then hope to see if there is a market."

If I had investors funding my startup business I too would opt for getting a usable product up and running ASAP too.   The robot products would be designed around as much "canned software" as possible. They would go out the door having been tested, and would hopefully perform as promised.

Would I want to do that as a hobbyist?  No.  But that may be not the answer for all hobbyists.  It all depends on what the individual hobbyist is interested in.  What are their goals?  What is it that they want to achieve?   They should certainly consider whether ROS would suit their goals and purposes too.

My own personal goals right now are to learn how things work in detail.  Loading canned software onto a robot to have it do things where I have no clue what it's doing is not my goal or purpose.   This may not be true for other hobbyists.   Each hobbyist needs to assess their own purpose and goals.

I'm currently taking a course on A.I. with Paul McWhorter.   Is this the quickest way to get into A.I.?   No, not at all.   There already exist programs that will do what Paul McWhorter is teaching.  Facial recognition, object tracking, etc.

The reason I'm taking these courses is because I want to learn and understand what goes on behind the scenes within these programs.  If I was running a startup business this would indeed be a waste of time.  The best thing to do with a start-up business is to just load up the facial recognition and object tracking programs that already exist, and move on from there.

So yeah, I agree with Mr. Tellez.  If you're going to manage a startup company, the best advice is to quite being a hobbyist.  Just get things done ASAP and don't worry that you don't understand how they work.  Just get the finished products out the door.  After you're rich and famous you can sit back and become a hobbyist again if you want. 😊 

DroneBot Workshop Robotics Engineer
James


ReplyQuote
frogandtoad
(@frogandtoad)
Reputable Member
Joined: 8 months ago
Posts: 365
2020-01-19 1:04 pm  

@robo-pi

Posted by: @robo-pi

The reason I'm taking these courses is because I want to learn and understand what goes on behind the scenes within these programs.

I agree with most of what you're saying... and I too always prefer to know how something works in detail, rather than follow someone else's framework... if you always follow, you will never invent or be a leader, and just follow like a good sheep!


ReplyQuote
Robo Pi
(@robo-pi)
Noble Member
Joined: 9 months ago
Posts: 1462
2020-01-19 1:05 pm  
Posted by: @frogandtoad

@casey

Hmm, that video was uploaded in 2011... I still don't see any robots of today with such capability, even if manually controlled?

I wouldn't even attempt to try to build a robot that could clean up a messy room until I had first built one that could get objects out and put them back away again from their proper places. 😊 

I'm still trying get my robot go in a perfectly straight line.  I'm working on adding an inertial guidance system to it as we speak.   It's going to have to learn how to get around in a perfectly clean and well-organized house before it can start learning how to put things away that aren't in their correct place.

In fact, I'm looking forward to the day when my robot can say to me, "This object doesn't belong here".  Then I'll be able to claim that I have succeeded in building the first artificial mom. 😎 

DroneBot Workshop Robotics Engineer
James


ReplyQuote
frogandtoad
(@frogandtoad)
Reputable Member
Joined: 8 months ago
Posts: 365
2020-01-19 1:19 pm  

@robo-pi

Posted by: @robo-pi

I'm still trying get my robot go in a perfectly straight line.

A perfect example and candidate for algorithm research! 🙂

Posted by: @robo-pi

In fact, I'm looking forward to the day when my robot can say to me, "This object doesn't belong here".  Then I'll be able to claim that I have succeeded in building the first artificial mom. 😎 

In a perfect world, indeed!

Every time I think about robots and how they should work, I always end up back in the same place (the foundation), that is, somewhere along the line, the following services must be present and linked at the hip:

1) Vision (most important in my view, and the greater precision the better)
2) Coordinate system for each moving component (arm, leg, lip, eye, etc...) mathematically linked to vision

I see the above as absolute minimum services, in coordination with a brain - I must be the first person to ever think of this, huh? 😛


ReplyQuote
Robo Pi
(@robo-pi)
Noble Member
Joined: 9 months ago
Posts: 1462
2020-01-19 2:22 pm  
Posted by: @frogandtoad

1) Vision (most important in my view, and the greater precision the better)

I keep coming back to humans.  There exist blind humans who function in the world pretty darn well.  In fact, I would actually prefer to have my robot able to get around the house without having to use an actual camera.  It's going to "feel" its way around with distance sensors and IMU.  Not unlike a bat.

This vision will come later enabling it to see in far greater detail and visually recognize objects.  

Posted by: @frogandtoad

2) Coordinate system for each moving component (arm, leg, lip, eye, etc...) mathematically linked to vision

Agreed.  If the robot has multiple moving systems they need to be linked together in a coordinate system if they are expected to work in coordination with each other.   This is true of even them most rudimentary robots.  If you have distance sensors on the right and left side of the robot, the robot had better know which side is right and which side is left.  

That's already quite important even with this little toy robot car I'm trying to get to go in a straight line.   If the robot didn't already coordinate it's right and left motors with the right and left orientation of the IMU, then it could hardly make the proper corrections

So yeah, the robot needs to what all its limbs are doing. 

DroneBot Workshop Robotics Engineer
James


ReplyQuote
casey
(@casey)
Reputable Member
Joined: 8 months ago
Posts: 328
2020-01-19 3:17 pm  
Posted by: @frogandtoad

@casey

Hmm, that video was uploaded in 2011... I still don't see any robots of today with such capability, even if manually controlled?

Missed seeing the age of the article.  Usually I take note of that. However from what I know it still holds. We are a very long way from a working robot beyond the current batch of vacuum robots that use vision and/or lidar.  A lot of hype to get funds and misleading videos making the robot look smarter than it is.

 


ReplyQuote
casey
(@casey)
Reputable Member
Joined: 8 months ago
Posts: 328
2020-01-19 3:26 pm  
Posted by: @robo-pi
Posted by: @frogandtoad

1) Vision (most important in my view, and the greater precision the better)

I keep coming back to humans.  There exist blind humans who function in the world pretty darn well.  In fact, I would actually prefer to have my robot able to get around the house without having to use an actual camera.  It's going to "feel" its way around with distance sensors and IMU.  Not unlike a bat.

A camera is cheap for a hobby robot whereas giving a robot the ability to use force feedback in its limbs and have anything approaching our sense of touch would cost a fortune and then we come back to the software problem of how to process all that information.

 

This post was modified 1 month ago by casey

ReplyQuote

Please Login or Register