Obstacle course

[Brian Higgins]

Which would work best for robotic sensors, a black box or a white box, as an obstacle?

Brian

Sent from my iPhone 12 Pro Max

[Michael Wimble]

If I understand the intent of your question, to create an obstacle that would be most easily detected, my first response would be that a better color would be one that doesn’t exist much in the normal environment, like fluorescent yellow or orange or purple. If your only choices are black and white, I would say that black is found everywhere in the environment where light is dim, while white is relatively rare. Using a rare color makes it somewhat easier to select something from the background. In my not-so-humble opinion.

> --
> You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/B79DFBEE-5F18-40DB-90BE-A2B9EE2457A4%40gmail.com.

[Chris Albertson]

If the sensor is a camera, I think what you want is something that is uncommon in the environment.   The color does not matter so long as it is uncommon.

If the sensor is a LIDAR or a "structured light" camera then light colors (such as whilte) are easier to detect as they reflect more energy back to the sensor.

But be careful if you are developing an obstacle detector that is to work in the real world.  You do not want a detector that only works for obstacles of one specific kind or color, mix it up a little.

--
You received this message because you are subscribed to the Google Groups "HomeBrew Robotics Club" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/B79DFBEE-5F18-40DB-90BE-A2B9EE2457A4%40gmail.com.

--

Chris Albertson
Redondo Beach, California

[Martin Spencer]

https://youtu.be/kn93BS44Das

> https://groups.google.com/d/msgid/hbrobotics/CABbxVHtZp%3D0xj5nw

> qyLyd4fis8Ed1RuxkQfOndqSW3VRsqufqg%40mail.gmail.com.

>

  

[Wayne C. Gramlich]

Martin:

In the future, when you post a link, please some text that explains what
the link is about. Bare links with no description are annoying.

Regards,

-Wayne

> To unsubscribe from this group and stop receiving emails from it, send an email to hbrobotics+...@googlegroups.com <mailto:hbrobotics+...@googlegroups.com>.
> To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/613DE9D0.25784.3DB054AB%40mspencer.geckosystems.com <https://groups.google.com/d/msgid/hbrobotics/613DE9D0.25784.3DB054AB%40mspencer.geckosystems.com?utm_medium=email&utm_source=footer>.

[dpa]

The real world makes some interesting obstacle courses.   Here's the basement of the building where I work, before it was remodeled and made more useful for humans but considerably less fun for robots...

https://www.youtube.com/watch?v=T47LkRYayLA

cheers!

dpa

[Brian Higgins]

I have measured 25 of the top obstacles the the blind face.  I will incorporate an average mass in my 4 obstacles.

Brian 

Sent from my iPhone 12 Pro Max

On Sep 12, 2021, at 10:25 PM, dpa <dav...@smu.edu> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/f3dd9154-273f-46e8-8442-2bf03887596an%40googlegroups.com.

[dpa]

Hi Martin. 

I thought I was responding to a query about robot obstacle courses, and suggesting that the real world supplies some good ones.   You response here is, you know, pretty condescending.  

Okay, how about this one, a bunch of trees and picnic table obstacles out at the lake:

https://www.youtube.com/watch?v=YXtZHJ-fCoM

or maybe this one, with a single enormous obstacle, a complex building:

https://www.youtube.com/watch?v=HlOPGIk0LGs

Both these robots respond to dynamic obstacles.  I'll dig around and find a video.

cheers!

dpa

On 9/13/21 4:25 AM, Martin Spencer wrote:

[EXTERNAL SENDER]

wall following is mostly reactive, not proactive or contemplative.  lack of dynamic obstacles makes avoids 1930ish easy. 

the video I posted is several years old.  this one is a few months old on an platform that is not man rated, servant class. eg., it can never be upgraded to have arms for loading the dishwasher from a counter top.  too small, insufficient power, etc.

great toybot!  been watching this low level of autonomy (wall following is a subset of line following) for 20+ years...

BTW, the subterranean DARPA robot challenge is to dev the AI mobile robot tech to do fully autonomous hunter-killer robots for urban environments.  DARPA's duplicity is no surprise-

as corporate policy we have done nothing to aid our military to lethalize our human quick sense/avoid of moving and/or unmapped obstacles AI robot nav tech.

[Brain Higgins]

I was looking for either the research PI wanted a certified obstacle course and I wanted to test the sensor in real life. I know you can stick a sensor on a computer but this is a human relying on a sensor not a robot.

Thanks for all the information 

Sent from my iPhone 12 Pro Max

On Sep 13, 2021, at 12:16 PM, dpa <dav...@smu.edu> wrote:



To view this discussion on the web visit https://groups.google.com/d/msgid/hbrobotics/78b12335-2bd6-4191-9c20-6738e88b06c6n%40googlegroups.com.