S2 Scribbler Finding object?

lestat_ega

Hi

I am programming for the S2 for a prototype of mine and I want my
S2 Scribbler to find some random points A,B,C...N where the number of points are
unknown in a open area and stay there for 15 mins in each one and and keep going
and finding the next point until all the points are founded. I have the
following questions:

1) What kind of sensor + receiver do I need ? for
each point (i was thinking something like radio frequency tags or magments) and
calculate the best route
2) How i can tell my algorithm to go back ? because
i will have a cable hook to it and i want the S2 to resolve if it is tangled and
keep searching the points


I hope that you can somebody can help me

GeorgeCollins

A couple questions:
- Are the points arranged in any way or are they distributed randomly? It can take a long time to find a point if you have no clues about which direction to head. Just getting a robot to cover every point in an area is a challenge in itself.
- Would there be any other navigational aids? A line could be very helpful in this case.


RFID would work for detecting points. I have never seen an S2 with an RFID reader attached but it should be possible. RFID stickers are unique and you need to be very close to read them.

A slightly cheaper and simpler alternative might be use the ColorPAL and have it detect colored stickers. There would be a lower number of points you could detect because of your possible color choices. But lets say you wanted to detect ten points on a hardwood floor. You could get ten bright primary colored stickers-- not brown obviously-- all their own color. The ColorPAL can detect shades very well. You could easily program the S2 to detect them and tell which ones it had visited and which one it hadn't.

It seems like it would be a fun problem to attack either way.

lestat_ega

They are distributed randomly and it is going to be in my back yard then I also have a pool, how i can ensure that my s2 wont drawn in to the pool and find the objects into the grass?

..............................

GeorgeCollins wrote: »

A couple questions:
- Are the points arranged in any way or are they distributed randomly? It can take a long time to find a point if you have no clues about which direction to head. Just getting a robot to cover every point in an area is a challenge in itself.
- Would there be any other navigational aids? A line could be very helpful in this case.








RFID would work for detecting points. I have never seen an S2 with an RFID reader attached but it should be possible. RFID stickers are unique and you need to be very close to read them.

A slightly cheaper and simpler alternative might be use the ColorPAL and have it detect colored stickers. There would be a lower number of points you could detect because of your possible color choices. But lets say you wanted to detect ten points on a hardwood floor. You could get ten bright primary colored stickers-- not brown obviously-- all their own color. The ColorPAL can detect shades very well. You could easily program the S2 to detect them and tell which ones it had visited and which one it hadn't.

It seems like it would be a fun problem to attack either way.

GeorgeCollins

I was thinking, how are you going to make sure it doesn't get stuck? I don't know what your backyard is like, but assuming it has grass, concrete, and a pool, there are probably some transitions in the ground surface that the S2 is going to get stuck on. Outdoor robots-- like the ones they use for Robomagellan contests-- usually have a lot better ground clearance. With that big of an area and those kinds of surfaces you probably would need to demarcate where the robot can go with white or black tape.

I would recommend trying the task indoors first.

erco

No easy task. GPS is accurate to several feet at best, occasionally tens of feet.

Prediction: this bot's on the bottom of the pool the first day. Within an hour.

Not trying to discourage, just to manage expectations. If it was easy, everybody would be doing it.