And the tutorial quest has started...

[Tsukubadaisei]

Ladies(I hope) and Gentleman,

my summer vacation has started and I just finished doing the parts shopping.

It is hard to believe that pile of junk will become a robot one day. BTW, that was about $340(36000￥).
The only thing left is the laser scanner that I will receive next week.
For those who don't know what I am talking about:http://www.societyofrobots.com/robotforum/index.php?topic=4011.msg31113#msg31113

Anyway I have two questions:
1)Can i post the tutorial as a pdf?

2)If not, what should I do in order to start writting?

The reason I want to post as a pdf, it is because it is more printable then the format used here. And if you are reading on the computer you still can use links like HTML. The downside is not be able to embed videos(but I can just post the links to them). And I love the two column format used in technical magazines.

Now here is my schedule for this contest:
1st week of July:Shopping(done)
2nd   "     "   "   :Circuit, basic programming and chassis(yes 70% of the robot will be finished by Saturday)
3rd   "     "   "   :Laser scanner programming(part1)
4th   "     "   "   :Encoder programming
1st   "     "  August:Laser scanner programming(part2)
2nd   "     "   "   :Line tracer programming
3rd   "     "   "   :Travelling
4th   "     "   "   :Travelling

In September there will be some updates but 95% of the tutorial will be finished but August.
I would like to use this thread to hear suggestions, extra things you want to see in the tutorial.

(about the mp3 feedback I was discussing in another topic, I think I will hack an very old Ipod shuffle I got. I was think about using my hacked psp but I dont think it will fit inside the robot ,but because I am using this robot for another real competition, music feedback will not be my priority until September/October(yes, after the due date). It will be a bonus I will add later)
 

[sdk32285]

Hey-
I have used the Hokuyu that you reference in the other link. Just be carefull since it is very sensitive to sunlight. and its range is severly decreased on wet dark soil.
http://www.societyofrobots.com/robotforum/index.php?topic=2372.msg15953;topicseen#msg15953

[Tsukubadaisei]

Hey-
I have used the Hokuyu that you reference in the other link. Just be carefull since it is very sensitive to sunlight. and its range is severly decreased on wet dark soil.
http://www.societyofrobots.com/robotforum/index.php?topic=2372.msg15953;topicseen#msg15953

I know. But I also know how to filter. That is one of the points of my tutorial. I will be talking about this URG sensor but my filtering methods are pretty much universal, so the reader will be able to apply this methods to the usual sharp PSD sensor and the robot will work outdoors even if it is a clear sky, hot summer day.
Thanks anyway for pointing out.

[Scheda]

I assume this is the laser you're using because it's referenced in your other post?
http://www.acroname.com/robotics/parts/R283-HOKUYO-LASER1.html

If so... 
That thing is EXPEEEENSIVE.

[sdk32285]

Tsukubadaisei-
What type of filtering methods are you looking at?

[Tsukubadaisei]

I assume this is the laser you're using because it's referenced in your other post?
http://www.acroname.com/robotics/parts/R283-HOKUYO-LASER1.html

If so... 
That thing is EXPEEEENSIVE.

I usually get lots of sponsors. Hokuyu is my sponsor this time. So it is free for me .lol

Tsukubadaisei-
What type of filtering methods are you looking at?

Physical filters(mostly covers) to increase the range and software filters used to to remove things that dont exist from the readings. I dont think those techniques have a name, but if they have you going to have until september when the tutorial is finished.

[sdk32285]

Hi
I found with the older scanner if it got dazzled the data it returned was complete garbage so the only reasonable thing that I could do was try to detect bad values and throw them out. I was using basic statistical methods where having an outlier could potentially bias the sliding average heavily. If you use a baysian approach than the bad values should be relatively harmless.
The other issue was that when things had a low reflectivity (wet dark soil) the maximum range of the sensor would drop down to as low as ~1m...

If Hokuyu is your sponsor try to get there new scanner http://www.acroname.com/robotics/parts/R314-HOKUYO-LASER4.html . Its only slightly larger yet has much better outdoor characteristics.

[Tsukubadaisei]

Hi
I found with the older scanner if it got dazzled the data it returned was complete garbage so the only reasonable thing that I could do was try to detect bad values and throw them out. I was using basic statistical methods where having an outlier could potentially bias the sliding average heavily. If you use a baysian approach than the bad values should be relatively harmless.
The other issue was that when things had a low reflectivity (wet dark soil) the maximum range of the sensor would drop down to as low as ~1m...

If Hokuyu is your sponsor try to get there new scanner http://www.acroname.com/robotics/parts/R314-HOKUYO-LASER4.html . Its only slightly larger yet has much better outdoor characteristics.

Let`s see then. I've got a senior friend of mine that uses this sensor for his research. He can get a range of like 3m even under rain. Plus my robot will not run over soil. The worst case possible would be wet asphalt. I will try to negotiate the new scanner, but the guy from hokuyu assured me that this model is more then enough for my tasks. Well he is THE guy, who am I to argue with him. Guess we will have to wait until the tutorial is ready. Anyway, thanks or talking about the baysian approach, I will see if that is the same method I know.

[Admin]

1)Can i post the tutorial as a pdf?

2)If not, what should I do in order to start writting?

I checked your member pages account, and it says you haven't logged into it yet. Let me know and I can resend you the loggin info.

I'll require you to post the tutorial like normal, but you can include a downloadable pdf copy on the same page.

[Tsukubadaisei]

That is good enough for me. I will do some progress first and then I will apply fo rthe account. Thanks.

[Tsukubadaisei]

If Hokuyu is your sponsor try to get there new scanner http://www.acroname.com/robotics/parts/R314-HOKUYO-LASER4.html . Its only slightly larger yet has much better outdoor characteristics.

I tried to get this new sensor but they refused to give me. The reason is that they want to promote this particular model I am using. They said this is one of cheapest model to produce and has a very high performance/price ratio. From what I could understand, that is the model gives them the highest profit. But that is fine, I shouldnt be complaining at all, after all, it is not everyday you get a 2 grand sensor for free.

[Scheda]

I usually get lots of sponsors. Hokuyu is my sponsor this time. So it is free for me .lol

Hmm, well I just might have to investigate getting me some of those after my first few bots!

[pomprocker]

so how are people suppose to follow this tutorial and build this robot if it uses a 2 thousand dollar laser?

that's out of my reach financially and I have a decent paying career in computers, and for most of the younglings in here that's going to be impossible.

[Tsukubadaisei]

so how are people suppose to follow this tutorial and build this robot if it uses a 2 thousand dollar laser?

that's out of my reach financially and I have a decent paying career in computers, and for most of the younglings in here that's going to be impossible.

For hobbyists, this tutorial has:

-line tracing (SOR doesn't have a dedicated tutorial yet)

-Rotary encoder and DC motors

-A "new" micro-controller(Renesas SH instead of the usual atmel's at mega)

-PID, mapping and filtering algorithms.

-other small stuff like team management(I actually have a team for this other contest in which I am competing), mechanical principles and tips for the chassis(I wont be posting a step-by-step tutrial for the chassis, they have already hundreds of them) and a few examples for main loops(applications).

The laser scanner is a bonus for professionals like ADMIN, sdk32285, JesseWelling, Dunk, Ai-bot and etc.... The laser scanner is like just 10% of tutorial. There are plenty of information in my tutorial you could use in your robot. And, who knows? Tomorrow you might be the one using a laser scanner. :3
Also the laser scanner can be sort of emulated using a servo and a sharp psd(of course, it is going to be much slower, less precise and with smaller range)

Finally I would like to make it clear my objective is not to have a cake recipe that people can just read and literally copy my robot. I want to build a tutorial that can be used as a reference for future projects. For example combine a sharp psd sensor tutorial from someone else with my line sensor and encoder tutorial and you get a racing robot. Another advantage you can completely ignore the laser scanner part and still have a functional and usefull robot in the end (if you really want to rip off my robot). My tutorial main points are the principles and techniques related to EACH element final robot. If you master the principles, you can build whatever you want.

"Give a man a fish and you feed him for a day. Teach a man to fish and you feed him for a lifetime."

[pomprocker]

niice, i've been wanting to do encoders, pid, mapping, etc...

[Tsukubadaisei]

Gentleman,

I have got good and bad news:
First the good news:
->My robot is FINISHED!!!!111!!!







->sdk32285 was talking about some sort of improved version of the laser scanner. It seems that I got this improved version (It seems they have put some kind of polarized infrared light filter inside for outdoors activity).

->Got a girlfriend.

->Got a new part-time job.

Now the bad news:

->Got a girlfriend.
->Got a new part-time job.
Therefore I dont have time to write the tutorial. I am sorry.
The fact is that I tried writting the tutorial 2 weeks ago but I barely finished the mechanics yesterday. I still have to write the electronics and programming, and the programming part is supposed to be at least 3 times bigger then the other ones. I figured out that with my busy schedule I wont be able to finish the tutorial in time so I will have to give up the SOR contest this time. However I have this other contest and after I finish it I am planning in upload here the schematics, blueprints and source code of the robot so if someone is interested s/he can try to build the robot by her/himself(I might help through PMs). Of couse I am not going to abandon the tutorial project but I will postpone it for now. By the next SOR contest(winter?) I am pretty sure the tutorial will be complete. I am really sorry if I hyped anyone, but getting a girlfriend was a surprise even to myself.

[airman00]

Now the bad news:

->Got a girlfriend.

Girls , what can you do about it ? 
Nice robot ,can't wait until it's documented

[sdk32285]

I tried to get this new sensor but they refused to give me. The reason is that they want to promote this particular model I am using. They said this is one of cheapest model to produce and has a very high performance/price ratio. From what I could understand, that is the model gives them the highest profit. But that is fine, I shouldnt be complaining at all, after all, it is not everyday you get a 2 grand sensor for free.

True that   

[Admin]

->Got a girlfriend.

You should write a tutorial on that, too.


(gasp! you choose a girl over a free Axon!? )

[Tsukubadaisei]

->Got a girlfriend.

You should write a tutorial on that, too.

It wouldn't be fair. Instant victory.