go_away

Author Topic: Distance sensor, model inquiry  (Read 3253 times)

0 Members and 1 Guest are viewing this topic.

Offline awally88Topic starter

  • Robot Overlord
  • ****
  • Posts: 212
  • Helpful? 0
Distance sensor, model inquiry
« on: April 05, 2008, 09:12:31 AM »
Hey guys,

I just had a question about using a distance sensor, I'm currently working on a group project at uni in which we build a robot which has autonomously go over a track with obstacles. For more information go here http://www.unsw.adfa.edu.au/acme/studentactivities/warman/index.html.

Anyway comes in two parts as there are two approaches that I have decided we can take to this problem. I'm not sure whether to use a sharp IR range finder or a sonic sensor to get the distance.

a) Will a either sensor be able to reflect of a 30 degree wooden incline? or will I not get enough reflection to make an accurate measurement?

b) as the track is 5m long and i need a pretty good accuracy with this what sensor would I use? I looked at the sharp IR more so than the sonic sensor but I found that the GP2D12 is only accurate 10-80cm and the GP2Y0A700K0F is accurate 1-5m. Is there anything which will encompass both of these ranges?

Thanks,
awally

Offline benji

  • Supreme Robot
  • *****
  • Posts: 832
  • Helpful? 0
Re: Distance sensor, model inquiry
« Reply #1 on: April 06, 2008, 03:21:13 AM »
there is the GP2Y0A02YK wich has a range of 20-150 cm

Quote
a) Will a either sensor be able to reflect of a 30 degree wooden incline?
i guess it will, although not sure, you just have to expirement with that.

Quote
I'm not sure whether to use a sharp IR range finder or a sonic sensor to get the distance.
it depends on the application,
sonars have longer range but ir's beam is much narrower so its better for mapping
good ol' BeNNy

Offline superchiku

  • Supreme Robot
  • *****
  • Posts: 953
  • Helpful? 5
  • cooll
Re: Distance sensor, model inquiry
« Reply #2 on: April 06, 2008, 03:23:46 AM »
go for the ir's coz like benji said ull find that mapping is easy with sharp ir coz of the narrower beam and also cost less than sonars if money is a criteria for u
JAYDEEP ...

IT AND ROBOTICS ENGINEER

"IN THE END IT DOESNT EVEN MATTER"

Offline awally88Topic starter

  • Robot Overlord
  • ****
  • Posts: 212
  • Helpful? 0
Re: Distance sensor, model inquiry
« Reply #3 on: April 06, 2008, 03:37:32 AM »

Hey guys,

just to clarify I'm not mapping specifically such that I'm storing the information, its more so I know how far I am from the obstacles and where to stop rather than actually mapping a course. It's a pretty simple course in terms of locomotion over teh majority of the track, the sharp IR is just so i know how far I am from goal states.

Offline superchiku

  • Supreme Robot
  • *****
  • Posts: 953
  • Helpful? 5
  • cooll
Re: Distance sensor, model inquiry
« Reply #4 on: April 06, 2008, 03:40:28 AM »
whatever still its ur choice sharp irs are inexpensive and are moderately affected by colour and light , sonars are costly and are not affected by anything unless u want totake out the transducer and receiver circuit.. ::)
JAYDEEP ...

IT AND ROBOTICS ENGINEER

"IN THE END IT DOESNT EVEN MATTER"

Offline benji

  • Supreme Robot
  • *****
  • Posts: 832
  • Helpful? 0
Re: Distance sensor, model inquiry
« Reply #5 on: April 06, 2008, 03:43:59 AM »
Quote
sharp irs are inexpensive and are moderately affected by colour

color?? what colors are they affected by?

Quote
its more so I know how far I am from the obstacles and where to stop rather than actually mapping a course
sonars work for ya, you can get advantage of their long range.(the beam is 30 degrees wide)
good ol' BeNNy

Offline superchiku

  • Supreme Robot
  • *****
  • Posts: 953
  • Helpful? 5
  • cooll
Re: Distance sensor, model inquiry
« Reply #6 on: April 06, 2008, 03:45:50 AM »
off course they are u dont think that ull get the same reading for a black coloured material and a white coloured material when they both are at a same distance , the reading will vary slightly with different colours in sharp ir
JAYDEEP ...

IT AND ROBOTICS ENGINEER

"IN THE END IT DOESNT EVEN MATTER"

Offline awally88Topic starter

  • Robot Overlord
  • ****
  • Posts: 212
  • Helpful? 0
Re: Distance sensor, model inquiry
« Reply #7 on: April 06, 2008, 03:47:48 AM »
Hey, I was thinking the sharp IR was the way to go, it's just my tutor at uni was adamant that I should use sonic sensors for this, i came to the conclusion that he's an idiot though so thats ok. Thanks for the help guys.

Offline superchiku

  • Supreme Robot
  • *****
  • Posts: 953
  • Helpful? 5
  • cooll
Re: Distance sensor, model inquiry
« Reply #8 on: April 06, 2008, 03:50:33 AM »
no he is not an idiot, he was telling u that for the obvious reasons i stated above as the problems with sharp ir ,

but all components have flaws, the sharp ir should be able to fullfill ur requirements
JAYDEEP ...

IT AND ROBOTICS ENGINEER

"IN THE END IT DOESNT EVEN MATTER"

Offline ajax spray n wipe

  • Beginner
  • *
  • Posts: 6
  • Helpful? 0
Re: Distance sensor, model inquiry
« Reply #9 on: April 06, 2008, 03:59:39 AM »
Why do you need to have a long-range sensor when you can just get a 10-80cm one and just have to detect the object when you get closer to it? Why do you have to detect it at 6 metres away?

Offline awally88Topic starter

  • Robot Overlord
  • ****
  • Posts: 212
  • Helpful? 0
Re: Distance sensor, model inquiry
« Reply #10 on: April 06, 2008, 04:09:16 AM »
superchiku - he is an idiot, a very different reason to this. He emailed me and he just recommended sonic as he has used them before and hadn't used sharp IR,

ajax - The 6 m was so i know when i get to the specific obstacle which is a certain distance (1.2 m) in from the start and it would have been a lot easier to do that that rely on timers which due to teh random nature of one of the obstacles has the very likely possibility of creating errors.

benji - the sensor will operate differently on different colours  as some colours reflect/absorb more light than others.

Offline superchiku

  • Supreme Robot
  • *****
  • Posts: 953
  • Helpful? 5
  • cooll
Re: Distance sensor, model inquiry
« Reply #11 on: April 06, 2008, 04:12:07 AM »
well if it will operate in different colours still it is reliable unless u use highly absorbing black or highly reflecting colours,

and abt the obstacle why dont u buy the other sharp ir which is sharp gpyXXXX some model it has a range of 10 to 150 cm well in the ur range of 1.2 m
JAYDEEP ...

IT AND ROBOTICS ENGINEER

"IN THE END IT DOESNT EVEN MATTER"

Offline awally88Topic starter

  • Robot Overlord
  • ****
  • Posts: 212
  • Helpful? 0
Re: Distance sensor, model inquiry
« Reply #12 on: April 06, 2008, 04:34:56 AM »
enclosed is an img of the track, sorry about the crappy paint job, we have to go from the left hand side to the right hand side over the ball pit.  I want to slow down for the ball pit and incline so I thought if i had a range finder that encompassed both of those distances then it would be good. instead I'll have to gauge distances from using data from the wall down the middle.

Offline awally88Topic starter

  • Robot Overlord
  • ****
  • Posts: 212
  • Helpful? 0
Re: Distance sensor, model inquiry
« Reply #13 on: April 06, 2008, 04:41:12 AM »

Also I have another question for everyone, does anyone know where in australia to get the GP2D12 or the GP2Y0A02YK? I havent been able to find them yet.

Offline superchiku

  • Supreme Robot
  • *****
  • Posts: 953
  • Helpful? 5
  • cooll
Re: Distance sensor, model inquiry
« Reply #14 on: April 06, 2008, 10:21:35 AM »
i dont know if it is available there or not , but u can always order them from the net
JAYDEEP ...

IT AND ROBOTICS ENGINEER

"IN THE END IT DOESNT EVEN MATTER"

Offline Admin

  • Administrator
  • Supreme Robot
  • *****
  • Posts: 11,660
  • Helpful? 169
    • Society of Robots
Re: Distance sensor, model inquiry
« Reply #15 on: April 11, 2008, 07:26:13 AM »
You don't want to use sonar in enclosed areas - your sonar will hit near-by walls and return false positives.

The only way around this is to use many sonar with a filter algorithm, but then a scanning sharp IR would be much cheaper . . .

 


Get Your Ad Here