PDA

View Full Version : [Project] Group project



ooops
07-30-2008, 06:00 PM
In the spirit of the "free stuff for bright minds" and DB's generosity, and also influenced from a RC forum where someone built an RC plane then shipped to someone in a different country to fly and then they shipped to a different person until it eventually goes around the world.
Would anybody be interested in doing a "community bot" where as one person provides the base and then ships it to the next to add "brains" then to someone else to add sensors and someone for programing and so on and so on until it is either done or way to heavy for the first base?
I am not so much speaking of a competition bot but rather something that would be cool for "Makers fair" or for the Trossen booth at Robo Games. Probably should sort out what it would/should do before getting too far along.
Granted it would be expensive to ship from point to point, if not right away certainly later on, and there would always be a concern of damage in shipping.
But, I would incur the shipping and supply the base to get it started.
I have two fairly robust 12v motors, a 2'x2' alum plate, a 12v gel-cell, a saber-tooth, and wheels to get things started. It would be a rolling chassis ready for the next step.

So, is anybody willing to step up?

Adrenalynn
07-30-2008, 06:33 PM
I'm not volunteering.

I think that's a really cool idea! A guideline or goal to meet with the device might help it along. So get it started!

Did I mention: I'm not volunteering.

metaform3d
07-30-2008, 08:45 PM
Cool idea! Kind of a physically "open source" robot. There needs to be a central coordinator and it seems to me that you just volunteered. I would be willing to contribute, given proper direction.


I have two fairly robust 12v motors, a 2'x2' alum plate, a 12v gel-cell, a saber-tooth, and wheels to get things started. It would be a rolling chassis ready for the next step. You might consider excluding the battery, since that will be the largest shipping cost. Most of us can lay hands on a 12v power source.

ooops
07-31-2008, 07:20 AM
Originally Posted by Adrenalynn>
Did I mention: I'm not volunteering.


Adrenalynn, I understand completely! But how about if we edit to say "I'm not volunteering - YET!"
Just knowing your there if we need you for troubleshooting would be a great contribution!



Originally Posted by Metaform3d http://forums.trossenrobotics.com/ambience/buttons/viewpost.gif (http://forums.trossenrobotics.com/showthread.php?p=16653#post16653)
There needs to be a central coordinator and it seems to me that you just volunteered.

Although my intent was to just start things off. I will be happy to coordinate the project and try to keep things moving along in some sort of reasonable order.

The first thing to sort out is what would be a cool project? Even if you can't or don't want to volunteer, please share any ideas and or suggestions. This is after all a community project.
The base is well, robust for a lack of a better description. So, it can and probably should be a bigger sized bot.
I expect that I can have it rolling by early next week. Pictures will be forthcoming soon of the pieces to give everyone an idea of what we are starting with.

So for now all we need are ideas for the project!

Electricity
07-31-2008, 11:55 AM
I'd be down!
I can provide a paint job or something..

ooops
07-31-2008, 11:58 AM
I'd be down!
I can provide a paint job or something..

Cool:)
Any suggestions on what it should do?

Electricity
07-31-2008, 12:22 PM
Cool:)
Any suggestions on what it should do?
Explore? That would be sweet. Like, you let it out your front door which it marks as home some how, then it just kind of wanders around, checking things out.
I imagine it rolling over to a dead leaf, picking it up, and examining it.

ooops
07-31-2008, 12:24 PM
Explore? That would be sweet. Like, you let it out your front door which it marks as home some how, then it just kind of wanders around, checking things out.
I imagine it rolling over to a dead leaf, picking it up, and examining it.

We would need to do a Johnny 5 voice: "Need input"

Electricity
07-31-2008, 12:27 PM
We would need to do a Johnny 5 voice: "Need input"
Yes! There we go.. Lets make a big ass J5esque robot. But keep the laser instead of a tool kit.. ;)

ooops
07-31-2008, 12:29 PM
Here is a picture of the base it measures 21.5" x 24" x 3/8"
http://forums.trossenrobotics.com/gallery/files/1/7/7/8/base-plate.gif

Here is one of the 12v motor's picture
http://forums.trossenrobotics.com/gallery/files/1/7/7/8/robot-12v-m.gif

ooops
07-31-2008, 12:36 PM
Maybe the first order of business would be to find someone with a milling machine to put this base plate on a diet!

Adrenalynn
07-31-2008, 01:01 PM
What you guys are describing is SLAM - Simultaneous Location And Mapping.

The most difficult problem in robotics today and the problem I'm most interested in conceptually.

4mem8
07-31-2008, 01:36 PM
M.mmm interesting concept, I would love to be involved, but have so many projects going at present I think my wife would kick me out if I took on another one, But if I sit in the back ground and see how it evolves there may be some electronic components that are needed and can help that way, I have truck loads. Nice idea.

ooops
07-31-2008, 03:22 PM
I would love to be involved

See that is the beauty of it. You can participate with ideas and suggestions at this point.
When at some point you see a need and have a solution then you can jump right in. Personally I think it would be awesome to have contributions from around the world!
A true global effort!

darkback2
07-31-2008, 04:28 PM
Ok...a huge part of me really wants to stay out of this one...like my wallet...But...I would be happy to contribute either time or equipment. How about making this robot a "Team Trossen" entry in a category in next years RoboGames...Like maybe advanced firefighting...

darkback2
07-31-2008, 04:29 PM
Also...how about we just build this thing in Crabfu's garage.

Electricity
07-31-2008, 06:01 PM
Also...how about we just build this thing in Crabfu's garage.
You can't! Thats where I live..

ooops
08-01-2008, 08:40 AM
Ok...a huge part of me really wants to stay out of this one...like my wallet...But...I would be happy to contribute either time or equipment. How about making this robot a "Team Trossen" entry in a category in next years RoboGames...Like maybe advanced firefighting...

As far as the wallet goes, I would suggest that most of us have "spare parts" lying around that are just waiting on a project or project idea. Here is a chance to clean up the workshop a little.

As for the advanced firefighter - sounds good, I will be excited with whatever would get the most members excited!



Also...how about we just build this thing in Crabfu's garage.


That would guarantee instant notoriety:)

Electricity
08-01-2008, 11:50 AM
We should consider starting a pay it forward style thread. Where someone posts up the stuff they are giving away, and stuff they are looking for. If you see something you need, just ask, and as long as you have some stuff to give away as well, you're good to go!

ooops
08-05-2008, 12:50 PM
Parts before-

http://forums.trossenrobotics.com/gallery/files/1/7/7/8/robot-parts-2.gif

Parts After -

http://forums.trossenrobotics.com/gallery/files/1/7/7/8/traci01.gif

http://forums.trossenrobotics.com/gallery/files/1/7/7/8/traci02.gif

http://forums.trossenrobotics.com/gallery/files/1/7/7/8/traci03.gif

Here is the base pretty much assembled.
The casters are the wrong sized and some parts have since been pulled and painted. All parts were spare in the bin ... does this qualify as a junk bot?
Payload capacity = 200+lbs I have confirmed that it totes 180+lbs around easily ... "Hill-billy seq-way" video to follow at first chance I get.

4mem8
08-05-2008, 01:08 PM
M'mmm Pretty hefty base, Should be a good platform for something interesting to build on.

Electricity
08-05-2008, 01:27 PM
Actually a ghetto segway would be kind of a cool idea..

4mem8
08-05-2008, 01:34 PM
Actually a ghetto segway would be kind of a cool idea..

Agreed, it would be a neat idea, I like it. What does everybody else think!!:cool:

Adrenalynn
08-05-2008, 01:38 PM
There was a cool one made from Plywood at the Maker Faire. Instead of resorting to caster/training-wheels, they actually did it the "right way" with a 2-axis gyro and accelerometer

ooops
08-05-2008, 02:54 PM
There was a cool one made from Plywood at the Maker Faire. Instead of resorting to caster/training-wheels, they actually did it the "right way" with a 2-axis gyro and accelerometer

But did they say "hold my bear and watch this..." just before hopping on?

ScuD
08-05-2008, 03:49 PM
But did they say "hold my bear and watch this..." just before hopping on?

I realize you're a robotic chicken, but that makes it even weirder to imagine you're running around carrying a bear :veryhappy:

Apparently Toyota now has a segway clone as well. Seems like the industry wants us to evolve into creatures without legs or something..

Electricity
08-05-2008, 05:21 PM
I laugh every time I see someone with a segway, its so stoopid looking. That being said, ours HAS to have a beer holder.. Maybe even a cooler of some sort.

sthmck
08-05-2008, 08:03 PM
Ever watch youtube videos of segway crashes? They are some of the funniest videos I have seen. In one a segway rep actually tells someone to try a flip it over, obviously think that it wouldn't flip. WRONG

ooops
08-06-2008, 08:37 AM
Well, if it is a segway style we are after, I won't have to spend $10 on new casters:)
I am not excited about a segway to ride, but a segway style bot would be sorta cool. I will need to relocate the motors and axles to the end of the base and that should pretty much provide for segway style operation, although the belts might not be a good solution since they function as a clutch and soften the motor torque. Hhhm, that might be an insurmountable problem with this motor/wheel config.
Thoughts anyone...

Electricity
08-06-2008, 11:46 AM
Lets not do a segway then? How about a high speed mobile refreshment stand. That way when your out for a run you can still have an ice cold martini..

ooops
08-06-2008, 02:32 PM
Ok, here is the video.
Let me splane some things first.
It was dark when I got it running and I know I won't be home before dark until next week so sorry, it is filmed in the dark.
If you are at work turn down the volume.
The big battery in the beginning was for traction. It is running off the smaller battery.

http://www.youtube.com/watch?v=n4iBKyR6XrU

ooops
08-06-2008, 02:39 PM
Oh, and by the way my 12YO posted the video to you tube for me, thus the credits, music, and Halo theme.

Electricity
08-06-2008, 06:38 PM
lol looks good. I was wondering what the haloness was about :P

LinuxGuy
08-07-2008, 07:49 AM
Apparently Toyota now has a segway clone as well. Seems like the industry wants us to evolve into creatures without legs or something..
We actually DO have to be careful that this does NOT happen to us. Due to having a bad balance problem (inner ear issue, I think), I've been using my power wheel chair when I go out alone. I will walk (and prefer to walk) when I am with somebody though. I do notice a definite weakening in my legs if I use my chair too much and don't get up and walk around some when I am near things I can grab in case of a balance problem.

8-Dale

ooops
08-07-2008, 06:01 PM
OK, I have realized that the current robot base is just way too big to send around to other members, which ultimately was the goal of the whole project. So I guess it is back to the white board for a bot that we can all "play with and contribute to" and send on to each other. Good news for me ... I have a big bot base to build something big on:) But don't worry it will still be a group project ... just without the shipping!

I am still very committed to the idea of a group collaboration.
Having thought more about it and still looking for ideas from you guys and gals here is what I suggest so far-
We will need a "sticky" on the TRC - oh wait - check that one off:)
We will need a "title or name" for the bot and the sticky something like TRCBOT, or TRaCi, or TRCGP, or ... you get the idea, and it probably shouldn't be left up to me!
It will need to be small enough to ship around.
It will need to be big enough to load up with sensors and other such goodies
It will need to have a "web presence" of some sort so ideally we can all either direct it around from a website or at the least see what it is seeing.
It should perform a useful function or task? - (more a question than suggestion)
It should move around. (what is the preferred method of locomotion...walking on 2 4 6 8 legs (or chopsticks:)), rolling on 2 4 6 8 wheels, tracks, or a combination of some or all?)
It will need some "brains" but we will need to decide on what kind after the size is sorted out.

Sooooooo, the project is wide open what should WE build?

Electricity
08-07-2008, 07:09 PM
Well, what do we want to do with it?
I think it should either be something unique, i.e. not a generic platform. OR we could design it to compete in the 09 robogames.

metaform3d
08-08-2008, 12:37 PM
How about TRAMP -- Trossen Robotics Autonomous Multiuser Platform.

Autonomy is a hard enough goal I don't think it's necessary to set one too far beyond that. It would be great to have a robot that could just be unpacked from its shipping container and set loose in a new environment. It would wander around, explore and generally not require a lot of supervision. When it gets back to its charging station it could post a rough map of where it went and some photos of anything interesting it found to its web site. Kind of a robot tourist & blogger.

Naturally if it did have any trouble getting around the current "host" would have to debug/repair it, or add new capabilities to handle the new situation.

ooops
08-08-2008, 12:45 PM
How about TRAMP -- Trossen Robotics Autonomous Multiuser Platform./quote]

I LOVE IT!!!

[QUOTE]
Autonomy is a hard enough goal I don't think it's necessary to set one too far beyond that. It would be great to have a robot that could just be unpacked from its shipping container and set loose in a new environment. It would wander around, explore and generally not require a lot of supervision. When it gets back to its charging station it could post a rough map of where it went and some photos of anything interesting it found to its web site. Kind of a robot tourist & blogger.


Sounds like the perfect project!
It can evolve to do other stuff if and when we see fit!
So what type of platform should TRAMP have?

Adrenalynn
08-08-2008, 12:54 PM
Wow. All I can say is wow. What an AMAZINGLY cool idea!

I'm thinking it needs a GPS board and a lot of non-volitile memory too so it can track its progress whenever it can get signal as it's winging/trucking/etc across the US.

I vote three wheel castor platform. Makes the SLAM side much easier.

ooops
08-08-2008, 01:15 PM
Wow. All I can say is wow. What an AMAZINGLY cool idea!

Credit to Meta




I'm thinking it needs a GPS board and a lot of non-volatile memory too so it can track its progress whenever it can get signal as it's winging/trucking/etc across the US.

Brilliant idea ... tracks around the globe


I vote three wheel castor platform. Makes the SLAM side much easier.
We have a three wheeler vote! Who else has an opinion?

Adrenalynn
08-08-2008, 01:32 PM
I also put in a vote for plastic base, and maybe something like the Lynx motors. Think light yet torquie so it can carry around the processing power. How about making it WiFi-based so that it can stream its data both to the hosts machine (and get control info and decision making there) as well as realtime streaming to a server? I'm willing to host the server-side in my datacenters.

I'm also willing to donate the ranging side. It's software incomplete, but my hardware is about ready to see the light of day... The software should be basically functional in another couple weeks, assuming I don't get a job by then. ;)

droidcommander
08-08-2008, 01:38 PM
That sounds awesome, I suggest a PC based platform that would make the wireless and web posting portion a snap.

I know I'm new here but I'm a professional web developer and I have an old PC that I can get working and donate to the project. I'd also volunteer for the programming for the web posting and anything else I can do to help.

I second the motion for a three wheeler as well


DroidCommander

metaform3d
08-08-2008, 01:46 PM
Agree on GPS and WiFi. That sounds exactly right.

4WD might be nice if we wanted to explore more outdoor environments or handle shag carpet, but it would certainly add to the weight and expense. Smooth indoor environments are hard enough, so two-wheel with caster design should be fine (to start with).

Adrenalynn
08-08-2008, 01:51 PM
I suggest we think about something like a mini-ITX. Much less expensive than a PicoITX, but still 12v aware.

Other donations we're going to need to think about: GPS board, Compass board, any other sensors we want in addition to my pseudo-LIDAR. We also have motor controller and sensor input board to think about.

The way I have my LIDAR built right now, it takes an image without the laser, then quickly lases the area and subtracts the original picture from the latest image, leaving [pretty much] just the laser behind. Then I use the original image to attempt video feature extraction as well. I suggest that if the feature extraction finds something it deems unique that that get tossed up on the server. Alternately, we could just push a jpeg of the original image to the site and do jpeg-picture-push streaming, or recombine/transcode them on the fly into mjpeg video at the server. Either is trivial for me - I have twenty years of code available for that video handling.

ooops
08-08-2008, 01:52 PM
That sounds awesome, I suggest a PC based platform that would make the wireless and web posting portion a snap.

I know I'm new here but I'm a professional web developer and I have an old PC that I can get working and donate to the project. I'd also volunteer for the programming for the web posting and anything else I can do to help.

I second the motion for a three wheeler as well


DroidCommander

First of all - Welcome dc!
Wow, welcome aboard the project:) We will hold ya to the web development portion!
How big is the PC?

Early polls show the three wheel platform in the lead!

Adrenalynn
08-08-2008, 01:54 PM
Agree on GPS and WiFi. That sounds exactly right.

4WD might be nice if we wanted to explore more outdoor environments or handle shag carpet, but it would certainly add to the weight and expense. Smooth indoor environments are hard enough, so two-wheel with caster design should be fine (to start with).

If we pick off-road wheels and an off-road caster we should be able to handle reasonable terrain and certainly shag carpet. We could go for eMaxx or tMaxx/sMaxx/*Maxx wheels in the front, trailing castor. The wheels will tend to pull the caster over whatever it hits. I've completed the Rubicon without winching in a 2WD vehicle. It's just choosing the right gear and the right tires. :)

Adrenalynn
08-08-2008, 01:59 PM
I'm pretty fond of these motors: http://www.roboticsconnection.com/p-51-dc-gearhead-robot-motor.aspx

They're not cheap, but the built-in encoders are sweet, and I think they're under-rated at 100oz-in. (In fact, mine rate 117oz-in @ 7v)

They also sport a standard 6mm shaft which is nice. Use the adapters from ServoCity and you can graft just about any wheel you want on them.

droidcommander
08-08-2008, 02:02 PM
The PC I have is a standard ATX Mobo everything minus the power supply fit into a 12x12x4 area. Plan was to replace the power supply with a dc to dc converter.

DroidCommander

Adrenalynn
08-08-2008, 02:04 PM
Pretty high current drain for battery operation though.

metaform3d
08-08-2008, 02:20 PM
So the LIDAR system provides the camera as well? That's pretty elegant. Can it be panned and tilted as well to point at things of interest, or does it have to be locked down to stay calibrated?

Adrenalynn
08-08-2008, 02:20 PM
Sorry for the double post...

You know, the more I think about it, the more I think the Stinger + Serializer is the perfect platform.

- It's competition proven
- It's aluminum, light and tough. They did a great job on the body of the tracked platform, the body of the Stinger looks even better
- It has the motors I was recommending
- The serializer is both bluetooth and wifi ready
- It has room to get processing and sensor packages on it if we build another deck
- It has room to put a couple Trossen Robotics stickers on it
- The aluminum won't muck around as much with the compass as would ferris metals
- It's reasonably priced for what we're getting

What we're doing is launching our own "mars rover" right here on planet earth. What if everyone took a photo of the 'bot in their environment, with GPS coordinates, and we could also build a 3rd-person photo album.

DroidCommander - how about a Google Map/Google Earth interface online to track its progress? I already have a key for some servers.

How about we give it its own website? I'll host.

How about allowing us programmers to update its "firmware" on the fly as soon as it hits a hotspot? That would rock. Continuous development remotely.

droidcommander
08-08-2008, 02:29 PM
Oh yea, we gotta do the google earth mapping. I have a couple buddies at work that do that all the time for their bike rides, shouldn't be that hard to automate it so a bot can do it.

Droidcommander

ooops
08-08-2008, 02:32 PM
Is this the one you mean = http://www.trossenrobotics.com/stinger-robot-kit.aspx

If so this looks awesome!!!!!!!!!



What we're doing is launching our own "mars rover" right here on planet earth. What if everyone took a photo of the 'bot in their environment, with GPS coordinates, and we could also build a 3rd-person photo album.



This is soooo cool!

4mem8
08-08-2008, 02:36 PM
cool platform for this project, It's starting to get interesting.

Adrenalynn
08-08-2008, 02:43 PM
You have to be one of the hosts, 4mem, just cause you're so "far out of the way"

ooops
08-08-2008, 03:16 PM
You have to be one of the hosts, 4mem, just cause you're so "far out of the way"

I would assit with the shipping just to see the "TRAMP tracs" on the google earth website of our little bot in NZ!!!!!

Adrenalynn
08-08-2008, 03:28 PM
No doubt! We definitely have to go with a SIRF-III chipset if we hope to have any signal in the belly of a plane, as well as the battery utilization... We will also want a high-gain antenna on it. In its low-power mode it should just log GPS tracks to NVRAM and nothing else. :)

ooops
08-08-2008, 03:43 PM
This is getting exciting!!!

ScuD
08-08-2008, 03:45 PM
Xan is from the Netherlands, right next to Belgium. we're somewhere mid-way between US and NZ? :veryhappy:

metaform3d
08-08-2008, 04:06 PM
How about we give it its own website? I'll host.Definitely! What I had in mind was a blog format, where TRAMP would be the main blogger (reporting his/her(?) position, photos, progress reports, etc.) and the builders would be "guests".


How about allowing us programmers to update its "firmware" on the fly as soon as it hits a hotspot? That would rock. Continuous development remotely.Absolutely. With detailed logs remote developers could play back events and find the coding error that ignored the left IR sensor and sent TRAMP tumbling down the stairs. Is there anything like a standard software platform for this sort of thing?

I was also hoping that non-technical visitors to the site could contribute. The robot could post photos with a question, like "is this a door?" and get answers in comments. Visitors could also propose missions to investigate unknown areas, and maybe get a chance to drive remotely (although that would require explicit permission by the host -- no one wants an internet snoop bot driving around their house).

Adrenalynn
08-08-2008, 04:13 PM
Awesome! Perfect! Just like a mars rover.

I think the coding platform will be dictated by hardware choice in this case. If we go with the Serializer we're most likely looking at developing our own code in C#. If we add on additional input devices, like an Axon, for example, we'll be looking at doing its firmware in either C or Assembler.

I don't think we'll find a pre-developed standard software platform to do what we want...

Oh - and I missed your question earlier: Camera PTZ - hmm. I've given a little thought to it but haven't tried any of the notions out yet. My concern is losing calibration since it's such a crucial sensor to navigation and mapping. I suggest we slap another cheap webcam on it for ptz, and keep the framesize/bandwidth down on the navigation camera and pretty on the vanity-cam.

metaform3d
08-08-2008, 04:27 PM
I don't think we'll find a pre-developed standard software platform to do what we want...OK. If we had a standard protocol whereby independent modules could synchronize with each other then they could be written in anything, at least for the high-level control software. Would this by a place to prototype some TRS concepts, or is that for solving a different type of problem?


I suggest we slap another cheap webcam on it for ptz, and keep the framesize/bandwidth down on the navigation camera and pretty on the vanity-cam.Sounds good. Ala Mars rover, it would be nice to have it on a boom -- the ground-level viewpoint gets a little tiring.

Can the LIDAR also be used in the scanning mode? If the robot found something particularly interesting it could make a hi-resolution 3D scan for the website.

Adrenalynn
08-08-2008, 04:31 PM
Yes, good point on the TROM/ROM and the applicability of a testbed, although its readiness is another question. I really need to drop everything for the whitepaper. It still haunts me.

I agree that putting the vanity camera up high is a good idea.

I hadn't though about switching the ranger over to a high resolution scanner on the fly. Let me give that one a little thought - there are complications involved.

[And with Metaform's idea, this thread just went to a 5-Star rating from me.]

Adrenalynn
08-09-2008, 04:39 AM
I hope no one objects - I took the liberty of registering "trampbot.org (http://www.trampbot.org)". Once the name servers zone update. it'll be up on a dual quadcore with two single quadcore failovers, 3gbit/sec bandwidth each on a 34gbit/sec fabric. If a different name is decided upon, I can host that one as well/in lieu of...

Now, we need a logo and a website. :) I'll install a blogging engine as soon as it comes to life for my dns, maybe sooner. Should we also have a calendar and project management app?


[edit] Did anyone think to ask TPTB (The Powers That Be) if they're OK with having their name tied to this, and under what conditions it'd be OK for us to use their Mark? Sorry - just that nagging business background thang. :)

Zone transfers are going now. I'm seeing it from Comcast and from Pacbell... I dropped a place-holder page up. I'll link it back to this thread.

I'll setup accounts on the site for anyone that needs 'em in the first go-round here...

metaform3d
08-09-2008, 03:55 PM
Did anyone think to ask TPTB (The Powers That Be) if they're OK with having their name tied to this, and under what conditions it'd be OK for us to use their Mark?If they don't approve the "TR" could just be changed to "Tele-Robotic".

If they do, perhaps they would like to contribute...

Adrenalynn
08-09-2008, 04:12 PM
Good point.

Are you ok with the domain name choice?

metaform3d
08-09-2008, 05:15 PM
Are you ok with the domain name choice?Works for me.

Thanks for hosting.

Adrenalynn
08-10-2008, 04:02 AM
That's the easy part!

We should hold a logo creation contest. There's a reason I'm a tech geek and not an art geek, ya know?

ooops
08-10-2008, 06:15 PM
That's the easy part!

We should hold a logo creation contest. There's a reason I'm a tech geek and not an art geek, ya know?
Holy cow, that logo is awesome! I want a T-shirt!!! Are you sure your not an art geek in a tech geek body?

Very good point on touching base with TP2B. I will do that Monday, some how since it is TRC born I assumed it would be a TRC logo'ed bot. It would only be right to see if they have concerns or have specific criteria. Maybe we can get them to sell/distribute the T-shirts;)

Adrenalynn, great thinking and great initiative to get the domain going!
I can't wait to see the tracks on the google earth map!!!

Scud, no worries, you will get your chance to host on the European tour!!!

Meta, the "tourist" part you mentioned has me thinking about finding ways to get TRAMP to cool places, the Great Wall of China comes to mind. Not early on mind you - but eventually.
Everyone, be thinking of where you would like to see TRAMP visit.

Thank you all for getting this a step closer to reality!
I am very exicted to take the next step this week and start the flow of parts!

Adrenalynn
08-10-2008, 07:12 PM
Thanks!

Sometimes it might be the opposite. Depends who has their checkbook open. I'm just a "geekDotMercenary". ;)

I fear to write what I'm about to write anyway - I might get laughed at by real art geeks...

I was laying in the hot tub kinda letting my mind float and conceived a logo that was "1950's NASA edu-tainment video meets Hitchhikers Guide". The inspiration was heavily influenced by the latter.

My design goal was to create something that could happily live with the Trossen logos. The font is a slightly heavier and more rounded version of the "I Build Robots!" logo, nothing jarring when you put them together. I envisioned the Trossen Star logo illuminating them - the shadow is painted such that the Trossen logo could take center stage above it and make sense. The silhouette of the robot on the "I Build" would then also work in the same fashion.

When I hopped out of the 'tub, it took me thirty seconds to find a free-rights/public domain globe (from NASA, actually...), which I then slapped on different "art papers" in Fractal Painter until I found one I liked - two minutes. The robot was next, I just sketched him directly using the tablet, then inked the lines with Photoshop. He's probably very similar to the sample in some cartooning tutorial I did a long while back. Can't for the life of me find where/when, although I did look. Otherwise I'd credit it]. I combined the two and put the ring underneath to support him and give him motion. It looked unfinished, so I went through some of the astrophotography photos I've taken over the years and slapped that starfield behind it. [Incidentally, behind the globe is a star going nova in Puppis last year. It was a nearly impossible shot to take in this hemisphere, I still can't believe S&T didn't feature that one - they've featured a bunch of my other astrophotography...]

Total time was a bit less than 15mins from tub to upload...

Anyway - just a proposal for a project-specific logo. I'm not married to it or anything, if something better comes along from a "real art geek" - we should be all over it!

----------------
Having it map part of the Great Wall would rock!
In geocaching we have "travel bugs". They're some trinket with a dog-tag on them, geocachers move them from cache to cache logging them in to the website as they go. Many want their picture taken with landmarks and uploaded to the site. In many ways this is like a travel bug - a travel bot. Really fascinating concept and very engaging. As an educational device as well. This gets me thinking that we want some other cool telemetry on it. Temperature? Humidity? Solar radiation? Soil/water content? Spectroscopy? Those are all really cheap and easy to build sensors [at the hobbyist level/resolution], but are environmentally/educationally interesting. How about a mammal-counter? A thermopile could be used to count the number of people larger mammals it sees.
------------
Where's our web engineer? I think Google Maps should be displaying the geographic location of those showing interest in the project so far. Our pre-site site should have a registration page where you can give your city or coordinates and get a pushpin on the map...

We could also have a "host family" control there where, for example, you could select The Great Wall and someone else could select "I can get it there" which would then result in the scheduling of travel...

Definitely have some webineering work [like that word? I just invented it. ;)] that could get going in parallel with the 'bot being sourced and built.

metaform3d
08-10-2008, 09:03 PM
Great logo! I'm fond of the retro space-age look myself.

The more sensors the better. Temperature, pressure, vibration and shock sensors would all be very useful for a robot that travels by mail. We would know where and when it fell off the DHL truck, and we could answer the question about how cold packages get at 30,000 feet.

We might be able to do some really interesting things on the web side with the raw data. If the robot takes enough photos of an area, we might be able to automatically combine that with the SLAM data to make crude 3D representations of the interiors. There's been some work done on that using web-archived tourist snapshots of famous buildings. By registering features from the different views they can back-compute the camera position and then reproject into 3D space. I'm at SIGGRAPH right now -- I'll see if there's been any improvements with that.

Adrenalynn
08-10-2008, 09:31 PM
I used to religiously go to SIGGRAPH - it's always been a great show - I miss it. I imagine it's mindblowing since it's been at least nine years since I was last there...

Yes, 3D projection was something I wanted to mention so you could chime in on. We'll have accurate positional data so certainly we should be able to get a pretty good approximation. The camera will also need to be calibrated at each startup [ideally], so we'll know what the optical deviation is.

We're starting to get enough of a sensor package to begin strongly contemplating a muxing scheme, otherwise we'll go broke on A/D...

I think we should also consider strongly selecting as low-current/voltage devices as possible. I suggest that a camcorder battery or two might be considered for off-line data logging. The online drive system can be less esoteric in power supply, but we are going to want to pack a lot of data logging into a very small power footprint. LiPoly camcorder or even cellphone batteries fit the bill nicely. LiPoly also has a really good shelf-life where self discharge is concerned.

Is this the point where we also start considering 2.5G and 3G/GSM (world-band) data uplinks?

Enjoy SIGGRAPH! Please keep us posted on what you learn in imaging and space projection! :)

Adrenalynn
08-10-2008, 09:54 PM
http://www.sparkfun.com/commerce/product_info.php?products_id=7917

Oh yeah. SIRF-III and worldband GSM modem, with external antenna connectors. up-to 57kbit/sec (2G) but it has a TCP stack w/FTP and SMTP built-in. Has an onboard PYTHON interpreter. Data, Voice, SMS, Fax - and can even handle a CMOS camera...

metaform3d
08-10-2008, 10:44 PM
SIGGRAPH had its heyday and has been on a decline for several years. The image-based rendering stuff has been one of the very few really innovative research areas lately. Since you can get all the papers and videos on the conference DVD there's not that much point going, apart from the parties.

For camera calibration we can just put some targets on the robot itself. Centering a couple of known targets would lock down the orientation, and if we add a color chart and gray ramp we can do automatic white-balance correction as well.

So, with that card the robot would have a cell phone? (Or be a cell phone...) Wacky.

Adrenalynn
08-10-2008, 11:12 PM
I've noticed historically that SIGGRAPH has its ups-and-downs too. Very early in the 90's it was sailing with SGI, Cray, Intergraph, Apollo, Sun, nCube, Ramtek, Cromemco... Then late in the 90's it took a downturn that seemed to trace Intergraph, Ramtek, Apollo, Cray, Cromemco, [and to a lesser extent, SGI] going away.

I wonder if they'll have their upturn again in a few years or if the nVidias and ATIs have inadvertantly driven the final nails into the coffin of high-end graphics by bringing super-computer rendering to the desktop?

Sorry - OT...

Yes, we can calibrate the camera onboard, we'd just need to make sure the calibration system was tough enough to survive what we'll be putting it through. Etched glass will be right out... ;)

Yup - ET will be able to Phone Home. Pretty wild, huh?

Running some rough numbers, if we sampled the GPSr every 15mins keeping it "warm" in between (cold-start is horrible for battery consumption with GPS), and attempted to phoned-home every hour, with near-realtime sampling on most of the rest of the proposed sensors (especially compass and accelerometer), we should be able to get 5-7 days from three or four BP-511 high capacity canon-knockoff batteries. They're only about $10ea from eBay suppliers, and they full-charge in about 40mins with my standard Canon charger.

asanist_13
08-11-2008, 07:48 AM
Hmm....
Nice member-to-member-collaboration project...
Hope I could lay my hands on it...
But sad to say...
Anyway, since most of our members live in "The Good-Old Wild Wild West" side of the planet, why not add a 5' solar panel?
You know, for extended "playtime"...

ooops
08-11-2008, 08:59 AM
Running some rough numbers, if we sampled the GPSr every 15mins keeping it "warm" in between (cold-start is horrible for battery consumption with GPS), and attempted to phoned-home every hour, with near-realtime sampling on most of the rest of the proposed sensors (especially compass and accelerometer), we should be able to get 5-7 days from three or four BP-511 high capacity canon-knockoff batteries. They're only about $10ea from eBay suppliers, and they full-charge in about 40mins with my standard Canon charger.

Just a thought on the "travel logging and Data". Since we will need a fairly custom case, I am thinking we can provide a battery pack within the case, which would provide for longer data logging and not add any footprint (space or weight) to the actual platform. Just plug the case in when it arrives to a host. Which raises a small question on power supply & plug issues. Here in the US we use 120v 60hz but that isn't the norm for the rest of the globe. Actually thinking maybe run all the charging through the travel case?
Similarly, we could provide for case based antenna's for the GPS and GSM Modem as well as any other sensors, just plug in the cord when packing it to ship.

At some point I wonder if there is an opportunity for a classroom "lesson plan" for following TRAMP around the globe???

I for one love the logo as is. Granted, a logo contest may grab some attention, but I don't think we can do better!

ooops
08-11-2008, 09:07 AM
Hmm....
Nice member-to-member-collaboration project...
Hope I could lay my hands on it...
But sad to say...
Anyway, since most of our members live in "The Good-Old Wild Wild West" side of the planet, why not add a 5' solar panel?
You know, for extended "playtime"...

asanist_13, hope you hop aboard the project.
The solar panel would be cool, but without the NASA style budget getting it to fit on the platform or in the case may provide a tough obsticle to overcome.

Adrenalynn
08-11-2008, 09:43 AM
I don't know if I'm sold on the battery-in-the-box notion. We're going to need all that power and functionality on the 'bot itself, right? The charger-in-the-box is an awesome idea, though! And that bred another idea - when the 'bot is in the 'box, it can cut its motor and logic batteries over to the datalogger as well. 4Hrs of motor runtime is like 9 days + of datalogger/gps-r time!

Ooops, since you're project lead here, don't you think you should start collecting the brainstorming into a project wishlist soon so that we can source the BOM and figure out the costs? :) I've just promoted you to project manager in one mighty swoop! ::tongue::

ScuD
08-11-2008, 09:47 AM
I can prolly provide a wcdma / gsm / gps module (it's what we build at work) but they don't have a sIRFIII chipset though.

In all honesty, I can't even tell wether the GPS reception is any good, but hey, it'd be free :veryhappy: or at least, free to the community...

ooops
08-11-2008, 10:13 AM
Ooops, since you're project lead here, don't you think you should start collecting the brainstorming into a project wishlist soon so that we can source the BOM and figure out the costs? :) I've just promoted you to project manager in one mighty swoop! ::tongue::

Hmmm...sounds like a call to get my stuff together:)
I will be sorting through the ideas and compiling a want/need list to be checked off.
Now is a really good time to add any new ideas!!!



I can prolly provide a wcdma / gsm / gps module (it's what we build at work)


Sweet, it's on the "need" list!!! Thank you:)

Electricity
08-11-2008, 12:17 PM
Lol so much for
I'm not volunteering.

I think that's a really cool idea! A guideline or goal to meet with the device might help it along. So get it started!

Did I mention: I'm not volunteering.
:D
I'm still in this, but the discussion is WAY over my head at this point. Just don't forget me when its time to paint the thing up!

droidcommander
08-11-2008, 02:09 PM
Hi everybody,
Sorry for being offline all weekend. I just got caught up on the thread and I'm excited that we have a web addr and logo started up. I'd be interested in getting the web site started if you have features/sections that need to be done. Also I'd need access to upload to the site, but that can wait itl we're ready.

I'll mockup some stuff and post it later today.

DroidCommander

ooops
08-11-2008, 03:09 PM
Hi everybody,
Sorry for being offline all weekend. I just got caught up on the thread and I'm excited that we have a web addr and logo started up. I'd be interested in getting the web site started if you have features/sections that need to be done. Also I'd need access to upload to the site, but that can wait itl we're ready.

I'll mockup some stuff and post it later today.

DroidCommander

DC, glad you are caught up:)
Adrenalynn has been bizzy.
We are really just waiting on the blessings from TP2B, and hopefully can go full speed ahead.
I am glad you are onboard!

Matt
08-11-2008, 04:16 PM
Just wanted to drop in and say that I think this project is pretty darn cool. When you guys have the initial order ready to go we will give a discount on it to support the project some. TRAMP is an awesome name BTW. Fits the idea perfect!

Electricity
08-11-2008, 04:26 PM
She does get around.. :p

asanist_13
08-12-2008, 06:58 AM
asanist_13, hope you hop aboard the project.
The solar panel would be cool, but without the NASA style budget getting it to fit on the platform or in the case may provide a tough obsticle to overcome.



Boy!!! I'd love to!!!
But, neh....
I'm better-off as a mere spectator than a co-builder...
For some 'bitter' reasons, that solar panel idea is all I can drop down...:veryhappy:

This might fit,though...
http://www.hobbyengineering.com/H2202.html

sam
08-12-2008, 07:35 AM
Wow! Your project is starting to form very fast! It looks very nice.

ooops
08-12-2008, 08:50 AM
Just wanted to drop in and say that I think this project is pretty darn cool. When you guys have the initial order ready to go we will give a discount on it to support the project some. TRAMP is an awesome name BTW. Fits the idea perfect!

Matt, Thank you!!! Your support and interest are truly appreciated!

I feel this project truly represents what the TRC is all about - bringing together a collaborative effort to build something cool, and then to launch it into the world.

So, if you have time and or expertise you would like to volunteer please do. (PM me)
If you have ideas to make it better please share. (post them here)
If you want to host a "mission" there will be a sign up for that soon. (on the official www.TRAMPbot.org (http://www.TRAMPbot.org) website)
If you just want to sit back and watch, that's fine too:) (Tune in as often as you can)

Look out world here comes TRAMP!

LinuxGuy
08-12-2008, 04:59 PM
I'm not interested in the hardware building, but I'd sure be interested in working on software. I've crammed some pretty good code into a small 13k flash space - an almost full subsumption engine. Unfortunately, I kept running out of program space for new behaviors and triggers. Give me a decent processor with decent flash storage, and I am sure I could do something good for this bot.

8-Dale

LinuxGuy
08-12-2008, 05:07 PM
http://forums.trossenrobotics.com/gallery/files/1/7/7/8/traci03.gif

Here is the base pretty much assembled.
This is a two wheeler then? I'm pretty good with software fort two wheeled robots, since that is what W.A.L.T.E.R. is (almost two years old now).

8-Dale

Adrenalynn
08-12-2008, 06:06 PM
That won't be the platform, but as the design stands now, yes it's a two-wheeled design. Windows-based, 2Ghz with a gig of RAM. Most likely the sensor preprocessing will be an AVR640 to start with, probably the Axon. 16 A/D, 3 serial ports + USB, I2C and SPI, 9 PWM channels, 22 servos. Two cameras (one dedicated to the 3D LIDAR), and a pretty good host of sensors - a lot of environmental stuff.

As designed (ie. assuming Ooops doesn't get cramps whilst writing the checks), it will have GPSr + GSM Cell in the first go-around, and will be updatable on-the-fly. It'll also have a channel of 802.11G for the high-bandwidth applications.

Adrenalynn
08-12-2008, 06:45 PM
Should P in "Platform" become "Project" ?

metaform3d
08-12-2008, 09:52 PM
Should P in "Platform" become "Project" ?Might not be a bad idea, although it could confuse the project with the robot itself. I was thinking that "Multiuser" could be "Mapping".

LinuxGuy
08-12-2008, 10:00 PM
That won't be the platform, but as the design stands now, yes it's a two-wheeled design. Windows-based, 2Ghz with a gig of RAM. Most likely the sensor preprocessing will be an AVR640 to start with, probably the Axon.
OK, I'm out then. I don't have any interest in doing development for Windows, sorry. I should have realized it would be a Windows based project before I chimed in.

8-Dale

Adrenalynn
08-12-2008, 10:07 PM
Might not be a bad idea, although it could confuse the project with the robot itself. I was thinking that "Multiuser" could be "Mapping".


Oopsie! I think I already made that ["mapping"] change without even realizing it. :o I think it just went through my head and out my fingers when I was doing the logo. ;)

As for the "P", I'm not sure what the confusion would be? The robot is the project is the robot, or are we thinking the project being larger than the robot - and is the logo for the project or robot? Sorry I'm not following. I might be slow tonight. ;)

-----
Dale - there's always the firmware side of the project. Fairly obviously the AVR640 doesn't run Windows. ;)

ooops
08-13-2008, 08:38 AM
OK, I'm out then. I don't have any interest in doing development for Windows, sorry. I should have realized it would be a Windows based project before I chimed in.

8-Dale

Dale, not so fast!!!
Your 2+ years of experience with WALTER could be invaluable to us! Also, I was hoping that you would jump in on some 3d CAD renderings!!! Once we have finally settled on the platform it would be great to show the world what TRAMP will look like even before we start assembly!


As designed (ie. assuming Ooops doesn't get cramps whilst writing the checks), it will have GPSr + GSM Cell in the first go-around, and will be updatable on-the-fly. It'll also have a channel of 802.11G for the high-bandwidth applications.

So far my hand is shaky but still functional:)
Isn't ScuD hooking TRAMP up with the GPS+GSM? Fingers crossed.

As for the "M" & the "P" I haven't a preference, all the possibilities discussed are cool and accurate! TRC community should chime in on which they like best!

Hopefully have some exciting updates later today!

metaform3d
08-13-2008, 12:09 PM
Oopsie! I think I already made that ["mapping"] change without even realizing it. :o I think it just went through my head and out my fingers when I was doing the logo. ;)You're right -- the logo does say Mapping, while the text at trampbot.org says Multiuser. I think Mapping is better -- more precise.


As for the "P", I'm not sure what the confusion would be? The robot is the project is the robot, or are we thinking the project being larger than the robot - and is the logo for the project or robot?Point taken. I was thinking of the actual hardware vs the web site, collaboration, etc. But you're right it's not really significant. I'm OK with either meaning of "P".

Adrenalynn
08-13-2008, 03:21 PM
I wasn't trying to make too much of a point (or be obnoxious) - I just wasn't grasping where you were going. I'm not married to anything but the TRAMP name - I LOVE that!

ooops
08-13-2008, 03:30 PM
Exciting update is in. Most of the "get it started" parts have been ordered!
I have a pretty cool (at least I think so) platform sorted out. Pictures to follow at some point in the future.
The list of "goodies" include -
Parallax motors and mounting kit
Sabertooth 25A controller
Sharp IR sensor - multiple
Sparkfun weather board
Vibration sensor
Voltage sensor
Mini-ITX
Axon
and more:)

So, we are one step closer! Albeit, a somewhat painful one for the American depressed card.
Anybody see anything we missed right off the start?

Comments / suggestions welcomed!!!

Adrenalynn
08-13-2008, 03:32 PM
Did you get the Sonar in there too? Presumably you did. You didn't mention the environmentals - keepin' that to yourself? :tongue:

ooops
08-13-2008, 03:39 PM
Did you get the Sonar in there too? Presumably you did. You didn't mention the environmentals - keepin' that to yourself? :tongue:

I was hoping you would share the environmental stuff:) You could splane it better.

I think I lumped the sonar in with the IR ... hope so anyway???
If not I will add it to the next order:)

Electricity
08-13-2008, 03:53 PM
Wow, so this is actually happening. Let me know if/when you need anything from me. The technobabble is still way above my head, but I'm willing to help any way I can.

ooops
08-13-2008, 04:00 PM
Wow, so this is actually happening. Let me know if/when you need anything from me. The technobabble is still way above my head, but I'm willing to help any way I can.

Are you still on board for the painting???

Electricity
08-13-2008, 04:14 PM
Of course! ;)

sam
08-13-2008, 04:57 PM
Wow! Your project is coming to reality! Congrats evryone who participated!

I love this thread, it advances fast compared to the other since there are so many users in on the sma eprojects, evryone has new ideas!

It's fun to read! I would love to help, but honnestly I don't know what I could do. :tongue:

Keep us updated!

metaform3d
08-13-2008, 08:47 PM
So, we are one step closer! Albeit, a somewhat painful one for the American depressed card. Anybody see anything we missed right off the start? Didn't realize you were already buying stuff. I was intending to/will offer my $25 Trossen credit from the contest if that will help.

ooops
08-14-2008, 09:00 AM
Didn't realize you were already buying stuff. I was intending to/will offer my $25 Trossen credit from the contest if that will help.

Meta, very generous of you! Hopefully we can take advantage of that on the next shopping spree when we realize what I forgot!
As with all new projects the parts shopping is the easy fun quick part ... assembly is the fun not so quick part ... and making it all work together is where hopefully everyone else gets involved.

Sonar was ordered, whew ... I was worried I missed it.

Once the parts arrive, and I have a few minutes, I will get a blog going to keep up with progress.



You didn't mention the environmentals
In addition to the sparkfun weather board http://www.trossenrobotics.com/store/p/5766-Sparkfun-USB-Weather-Board.aspx which will log: current pressure, humidity, ambient light level, and temperature. Hopefully we can add a few of these other way-cool environmental sensors: http://www.futurlec.com/Gas_Sensors.shtml : CO, CO2, O3, and AQ. So, if we can put this all together we should be able to impress Al Gore with the array of environmental info we can capture and upload!

sam
08-14-2008, 09:23 AM
The sensor senses a LOT of things!

I find it pretty interesting to have a robot that can "see" all of these things. Have data logs with where it is compared to what it sees on various days.

metaform3d
08-14-2008, 09:36 AM
OK, I'm out then. I don't have any interest in doing development for Windows, sorry. I should have realized it would be a Windows based project before I chimed in.This was what I was concerned about WRT having a common software platform. Once we pick C on Windows -- or Linux for that matter -- we have excluded some subset of people from being interested or able to participate.

Perhaps we could write high-level code in a cross-platform scripting language. Python is relatively easy to embed, for example. Maybe that just pushes the problem over to a different language that people would have to learn, but it does mean that anyone could write code for the platform without needing a Windows compiler environment.

ooops
08-14-2008, 09:50 AM
The sensor senses a LOT of things!

I find it pretty interesting to have a robot that can "see" all of these things. Have data logs with where it is compared to what it sees on various days.

The cool part will be "seeing" all the info logged as it is shipped around the world. As Meta mentioned earlier, Just how cold does it get in the belly of an airplane at 20,000'? Or how does the Air Quality here in the mountains of North Carolina compare to say the air quality in New York City, Las Vegas, Silicon Valley, Northern California, Southern California, Europe, New Zealand ... all the places that TRAMP will go.
As Adrenalynn mentioned this will be the "On Earth" version of the Mar's Rovers, roving around our planet (with the help of hosts) collecting and logging data, mapping it's environment, and sending back to the website cool video and data of all it's adventures ... just without the NASA budget (I hope!)

ooops
08-14-2008, 10:23 AM
This was what I was concerned about WRT having a common software platform. Once we pick C on Windows -- or Linux for that matter -- we have excluded some subset of people from being interested or able to participate.

Perhaps we could write high-level code in a cross-platform scripting language. Python is relatively easy to embed, for example. Maybe that just pushes the problem over to a different language that people would have to learn, but it does mean that anyone could write code for the platform without needing a Windows compiler environment.

Meta, is there a chance that you would head up the software portion of the project? Obviously I realize that is asking a lot, but I don't know anyone more qualified! These things always work better when one person has the "clear view" of the direction to go.
It would seem to me the easy part will be the motion and collecting the data, packaging it for up load seems a bit "trickier", as does the environmental mapping.
Please share your thoughts on all of this.

Adrenalynn
08-14-2008, 10:46 AM
Yes, there needs to be a software architect there. The packet structure needs to be architected, chunks of which are dependant upon the hardware and firmware though.

I would hate to see Ooops spend all that money on a 2Ghz low-power machine so that we could immediately make it 50% slower than it needs to be by specing Python on the bot. That seems wasteful.

If someone really wants to participate - the microcontrollers aren't windows based, the server isn't windows based (Linux RH Enterprise), and you don't need a windows environment to cross-compile to windows. [shrug]

---------

You know - I just had a thought related loosely to the environmental monitoring stuff.

NASA flies various experimental packages. If some qualifying university wants to know how mice do the dirty in zero gravity, well, NASA is likely to fly that package. We could do the same thing. Short term and long term "experiments" designed by others here on the forum and "flown" by The TRAMP...

Thoughts?

ooops
08-14-2008, 11:12 AM
So, that would be a "secure uploading of specific data"?

Adrenalynn
08-14-2008, 11:54 AM
I think that would need to be an option in that paradigm if we were getting universities involved. A period of time that allowed someone to publish before the information because public - but after that period of time the data co-belongs to TRAMP and we could do with it as we see fit. Or they maintain academic publishing rights, or something along those lines. That's a great question, honestly I hadn't really thought through all the ramifications.

I'm really excited myself about the complete environmental package. I think it would be really fascinating to overlay air-quality snapshots on Google Earth. We wouldn't have long-term and controlled data so the scientific value is pretty low - but the public interest value is really high. And anything involving the environment at this point is a press hotbutton of immeasurable intensity. ;)

sam
08-14-2008, 12:08 PM
You could rate cities air quality and do a list, see which one is best. :wink:

It would be much more interesting, having data for each place it goes too!

droidcommander
08-14-2008, 12:54 PM
On the topic of programming I'd like to formally suggest Java. Java is cross-platform and with any hardware there is usually a c/c++ API that can be easily wrapped using Javas' (Java Native Interface [JNI])

Practically this also helps with modularity of the code, we can discuss the API that we will write to for each module and each contributor can code that piece. once finished compiling all the source into a single Jar that comprises all the code is nearly as simple as zipping up the compiled source.

Additionally, Many of the API's that we will be using for posting to the Internet (Blog, google earth, general web content, video, images) have Java interfaces that can make using these easier.

I must admit that I'm not familiar with developing for some of the microprocessors but if just the main part were in Java any of those libraries (C/C++ I would assume) could be called natively from Java and those portions written in C/C++

P.S. I have built a web site on a local server and hope to have it up soon once I get ftp access. Then you can all chime in on the design.

My two cents
DroidCommander

Adrenalynn
08-14-2008, 12:57 PM
login pm'd to you.

As far as Java goes, I'd like to get my vote against. Slooooooow with no useful support for microcontrollers. I would never consider firmware with Java myself. But that's just me.

metaform3d
08-14-2008, 12:59 PM
I would hate to see Ooops spend all that money on a 2Ghz low-power machine so that we could immediately make it 50% slower than it needs to be by specing Python on the bot. That seems wasteful.Interpreted code would be slower, but there could be advantages in terms of flexibility. Python has an immediate mode, where you enter commands or queries and get responses back on the next line. You could write a short loop, right in the web page while the bot is running, which scans a sensor across a range of positions and reads out the values into an array. Then enter another command to plot it, for example.

I assume closed-loop control would mostly use microcontrollers rather than the main CPU, and computationally intensive operations like vision or Bayesian analysis would be compiled, but a lot of high-level control and decision-making for a mission could be handled easily by interpreted code on a fast computer. It's just a thought, and is something that can readily be added later. It doesn't require any more planning up front than would be needed for all good software architecture.

metaform3d
08-14-2008, 01:06 PM
Meta, is there a chance that you would head up the software portion of the project? Obviously I realize that is asking a lot, but I don't know anyone more qualified! These things always work better when one person has the "clear view" of the direction to go.I'm afraid I don't have a lot of experience with the types of problems that will be required to get this up and running. Serial communication, TCP/IP, network protocols, web backends -- these all remain a bit mysterious to me. Once we get to image processing and 3D visualization I'm ready to help!

Adrenalynn
08-14-2008, 01:13 PM
As long as it's just scripting, it probably doesn't matter. I just don't want us encoding video or doing trig in python/java/insert-interpreted-language here. ;)

droidcommander
08-14-2008, 01:30 PM
I agree, I wouldn't want to do video processing on Java either. I think that should be handled by a compiled library. But as for bringing it all together and gluing the disparate portions to each other I think that Java would work well.

P.S. The web site is up for those that want to see.

DroidCommander

JonHylands
08-14-2008, 02:15 PM
Doing trig in an interpreted language, as long as it is running on a decent piece of hardware, is a no-brainer. Any decent language (like, say, Smalltalk) does all the floating-point math in primitives that are written in C, and use any hardware support the processor has.

ooops
08-14-2008, 03:08 PM
Wow, great work guys things are moving forward:)
DC, thank you!!!
Adrenalynn, Thank you again!!!
Website for those that don't recall: http://www.trampbot.org/
I wonder if we should link the parts in the "parts list" back to the Trossen catalog?
Maybe Matt will want to do a "Tramp package" kind of a kit of parts thing ... no royalties required:)LOL
If the blogs are done well, they could be easily be turned into tutorials on various aspects of robot building/programing.

As far as the language/software conversation, I don't feel I am qualified to participate in the conversation that alone make the decisions on what to use. That is why I was hoping to hook Meta (or anyone willing to volunteer) into the job of making the calls on how best to proceed.

Adrenalynn
08-14-2008, 03:09 PM
Is it my imagination or is the project about to become eternally bogged in the religious debate of "which language is better"? :)

I retract my previous disagreement. I now support our implementing a BASIC interpreter under a Java interpreter under a Smalltalk interpreter written in Python that was compiled against ADA designed in COBOL but implemented in some combination of FORTH and PERL. We should probably implement a custom operating system in APL to run it on, and a JIT compiler for it written in SNOBOL that itself was built in LISP with any floating point converted to integer in ProLog. We should then use that to develop a realtime motion prediction algorithm for a hybrid of wavelet and DCT video compression with a required input aspect ratio of exactly 21.976:12.002. We'll stream it over Token Ring after extracting both the tokens and the ring and then splicing the coax into a 6 wire phone cord which we'll wrap into a ball and extract the emitted RF from using a tube-driven front end, a transistor driven IF, and a discreet IC amplifier.

Whew. I think we've now managed to exclude every - err - I mean - no one.

ooops
08-14-2008, 04:06 PM
Is it my imagination or is the project about to become eternally bogged in the religious debate of "which language is better"? :)

Well, so far I think we are on an amazing pace. We can afford a little time for discussion on languages and software, especially if it brings more folks into the project. Heck, debating such things is fun and for us "sudo-programmers" enlightening.
The thought I have is that the language chosen should be common. Someone that is willing in the beginning to code may not have enough time later on to devote and so someone else may need to add, enhance or change some of the code. If we use a common language like C/C++ ... that shouldn't be a problem, if we go with something less popular or more specialized then we may have trouble recruiting a new volunteer.

droidcommander
08-14-2008, 04:09 PM
LOL!!

Wow that's quite a post, but you forgot C# and VB and...

Really, I'm fine with whatever is easiest. And since I haven't built many bots with these specs I'll tip my hat to those with more experience.

Thanks
DroidCommander

Adrenalynn
08-14-2008, 04:16 PM
LOL!!

Wow that's quite a post, but you forgot C# and VB and...

I did so deliberately. :)

metaform3d
08-14-2008, 07:24 PM
A group software project requires some kind of decision about platform and language. I'm most comfortable with C/C++ on MSWindows, so that default is OK by me. I was just wondering if there was an architecture or approach that would be more flexible in terms of openness.

As an example -- and this is only by way of illustration because I'm familiar with it and not a proposal by any means -- COM is a well-defined binary standard. This allows COM objects to be created in C/C++, the .NET variants, Perl and Python, probably Pascal and Smalltalk for that matter. To a lesser extent they can be instanced over the network.

On another note, I've only ever written code that I compile and run on the same machine. (In fact at a job when I had to download code to a separate machine to test it I wrote a simulator so I could run it on my workstation. My simulator was so popular it spread all over the company.) I'm very curious to see how it will work to develop over the internet with a remote robot as the target platform.

ooops
08-15-2008, 09:10 AM
The big brown truck should be showing up soon. I think I will hang out at the loading dock and wait for the goodies to arrive:)
Pictures later today!

ooops
08-15-2008, 11:30 AM
First of many boxes to arrive!
:)"Some Assembly Required" :)

Adrenalynn
08-15-2008, 01:11 PM
Beach'in! Grats on the shipment!

ooops
08-15-2008, 02:05 PM
I forgot this:

sam
08-15-2008, 02:14 PM
Cool! You got your pieces!

Those looks like sturdy motors you got! I can't wait to see when you will start building it!

droidcommander
08-15-2008, 02:18 PM
I created a Blog for the TRAMP-Bot on Blogger. Check them out via the TRAMPBot website.
http://www.trampbot.org/blog.php

Pictures of the TRAMP in pieces are posted there as well

Thanks
DroidCommander

ooops
08-15-2008, 02:31 PM
I created a Blog for the TRAMP-Bot on Blogger. Check them out via the TRAMPBot website.
http://www.trampbot.org/blog.php

Pictures of the TRAMP in pieces are posted there as well

Thanks
DroidCommander

Great stuff - Awesome start!!!

metaform3d
08-15-2008, 02:57 PM
I looked into some of the literature on projecting images into a 3D virtual space and there are definitely some interesting things we could do. Especially since we control the platform we can send it back to fill in blank spots in the coverage by taking new photos. All of it, however, is very compute-intensive.

One thing we could do relatively easily would be QuickTime VR. (See here (http://anti-matter-3d.com/Freelance_QTVR_Services.html) or search the web for samples.) Panoramas can be generated automatically provided you have a set of photos that encompass a 360-degree view of the environment. The robot could go to the center of a room (or a set of other waypoints) and do a scan, taking 10-20 photos at each location. These can then be stitched on the server and posted, and with links form a virtual walkthrough.

The ideal way to handle this would be to have a camera mount that can swivel a full 360-degrees, plus 90-degrees or so of pitch, preferably with the focal plane of the camera at the center of rotation. Can we build/buy something like that? Alternately the robot could turn in place although that might not be quite as accurate.

A lot of pixels would also be nice, since we can take more photos and just upload strips from the center of each, which results in less distortion. What resolution camera will we be using?

ooops
08-15-2008, 03:19 PM
I looked into some of the literature on projecting images into a 3D virtual space and there are definitely some interesting things we could do. Especially since we control the platform we can send it back to fill in blank spots in the coverage by taking new photos. All of it, however, is very compute-intensive.

One thing we could do relatively easily would be QuickTime VR. (See here (http://anti-matter-3d.com/Freelance_QTVR_Services.html) or search the web for samples.) Panoramas can be generated automatically provided you have a set of photos that encompass a 360-degree view of the environment. The robot could go to the center of a room (or a set of other waypoints) and do a scan, taking 10-20 photos at each location. These can then be stitched on the server and posted, and with links form a virtual walkthrough.

That would be brilliant!


The ideal way to handle this would be to have a camera mount that can swivel a full 360-degrees, plus 90-degrees or so of pitch, preferably with the focal plane of the camera at the center of rotation. Can we build/buy something like that? Alternately the robot could turn in place although that might not be quite as accurate.

That should be easy assuming a fairly light camera. Agree, turning the bot is not the best solution.


A lot of pixels would also be nice, since we can take more photos and just upload strips from the center of each, which results in less distortion. What resolution camera will we be using?

Well, this is one area that I haven't spent enough time on.
Cameras - We all know we will need one/some so, hopefully someone has a few suggestions.
I envision just a simple "webcam" for the on the fly stuff.(probably not realistic for what we are doing or wanting to do) Also, an IR type camera would be cool, even if it were just still pictures from it. I know some here have extensive backgrounds in photography and cinematography, so please chime in!

We want Imax quality on a Kodak budget from a camera that weighs less than the average computer mouse! Seems simple enough:)

metaform3d
08-15-2008, 03:30 PM
Cameras - We all know we will need one/some so, hopefully someone has a few suggestions.I'm sure we can figure something out. Probably a lot of cameras can handle both tasks: low-res video for the webcam, high-res stills for archiving. For quality images resolution is ultimately less important than optics.

Another feature that would be very handy would be to have control over the f-Stop. This would allow us to create high dynamic-range imagery -- HDRI. HDRI images are nice because you can adjust them in post to correct for bloom and other unwanted effects. An HDRI panorama acts as a "light probe" and lets you composite or render any virtual object as if it was in the same environment. It's very cutting edge.

Adrenalynn
08-15-2008, 03:32 PM
It's all dollars. I think the Logitec Quickcam Pro 9000 is probably a really good choice.

Real glass Carl Zeiss-made optics, real 2mpixel (720 lines), USB2. Auto focus, auto color, color enhance, digital light enhancement. Under $75 if you shop around. It's small and light and it's Logitec = Reliable name.


"The settings that can be controlled through software include pan/tilt, zoom, auto/manual focus, face tracking, brightness,
contrast, color intensity, and white balance. Aside from those main video controls there are audio controls, the RightLight adjustment (by default this is on auto), advances settings, and then controlling how the cam works with your system (where files are stored, default applications, etc.)."

For navigation I'll be donating a cheapie camera, USB 1.1, low resolution, good optics, black and white (for contrast. I'm thinking about throwing a bandpass filter on it too)

metaform3d
08-16-2008, 09:40 PM
Adren: Funky form factor. Given what ooops wants to do I wonder if we could mount it vertically. Then we'd have 1200 wide by 1600 high, which would be much better for QTVR. We'd need to do on-the-fly rotation for web video though.

ooops: I saw the pictures you uploaded! Have you got something to share with the group?

Adrenalynn
08-16-2008, 09:56 PM
Yeah, it's a standard video resolution rather than photo resolution. Higher resolutions are interpolated. I could be wrong, but I've always been a fan of moving my interpolation off to servers. Infinite processing power, infinite bandwidth - and you're not using the limited bandwidth at the platform to send pixels that are artificially created...

ooops
08-19-2008, 09:43 AM
ooops: I saw the pictures you uploaded! Have you got something to share with the group?

Everyone wants to see the baby but no one wants to hear about the labor pains. So I will limit the update to the good news:)
I have found a great platform solution. I am having some custom work done on the chassis at the moment, the machine shop says by the end of the week they will have their end done, and I can get some pictures up, hopefully of TRAMP running around via RC. Then it will be time to start placing some sensors, and sorting out the next step. Slowly progress is being made!

droidcommander
08-19-2008, 11:14 AM
ummmm..... :( TrampBot.org is no more, tis disappeared from my DNS server

ooops
08-19-2008, 11:51 AM
I still see the website???

Adrenalynn
08-19-2008, 02:08 PM
Ditto that, as I mentioned in my PM.

Adrenalynn
08-19-2008, 02:29 PM
Checking various name servers online, all appears well:

Server: tethys.ringofsaturn.com
Address: 71.164.232.42#53

Non-authoritative answer:
www.trampbot.org (http://www.trampbot.org) canonical name = trampbot.org.
Name: trampbot.org
Address: 75.125.238.226

---
[70.84.161.11] returned a non-authoritative response in 16 ms: Answer records

nameclasstypedatatime to livetrampbot.orgINNSns2.adflix.com86400s(1d)trampb ot.orgINNSns1.adflix.com86400s(1d)

ooops
08-20-2008, 11:29 AM
Update - Shiny new parts for TRAMP
Axon
ITX
Some Assembly required:)

metaform3d
08-20-2008, 09:58 PM
Oooo, nice! I've been looking at the Axon lately; it's an interesting device. What's it intended for?

Adrenalynn
08-20-2008, 10:19 PM
It's going to be 2/3 of our analog input and sensor preprocessing. 16 analog inputs, three com ports + USB, up to 22 servos. It'll handle the GPS too.

They're pretty spiffy. I already have a bunch of libraries written for it...

ScuD
08-21-2008, 02:02 PM
It'll handle the GPS too.



Does this mean you've allready got a gps device?
I told Ooops I could provide a combined GPS/GSM/WCDMA device, but if you already have gps you'd only need gsm/wcdma, right?

Adrenalynn
08-21-2008, 02:30 PM
No GPS I'm willing to part with, but I'm assuming it will be handling GPS/GSM at the least. Some reason why it won't?

ooops
08-21-2008, 02:46 PM
Does this mean you've allready got a gps device?
I told Ooops I could provide a combined GPS/GSM/WCDMA device, but if you already have gps you'd only need gsm/wcdma, right?

We need it - GPS/GSM/WCDMA - we need it!!!!

Update on the chassis, the machine shop says they are working on it as I type, I hope to be able to pick up the finished product tomorrow!

Adrenalynn
08-21-2008, 03:32 PM
Which won't fit [the chassis] or look anything like you expected...

On the plus-side, you have a 50/50 chance of picking up a new ferrari vs. a lump of unrecognizable aluminum. Good luck! :tongue:

ooops
08-22-2008, 08:05 PM
Which won't fit [the chassis] or look anything like you expected...

On the plus-side, you have a 50/50 chance of picking up a new Ferrari vs. a lump of unrecognizable aluminum. Good luck! :tongue:

No chassis, although I was promised in person that it WILL be done on Monday.
They seem to be way-over engineering this thing, a simple chunk of alum would do ... we will see what we get. Hoping for the Ferrari, but time will tell.

Adrenalynn
08-22-2008, 09:03 PM
As long as it doesn't cost you a quarter million dollars. ;)

droidcommander
08-27-2008, 10:38 AM
A HUGE thanks goes out to Jodie for helping me with the TRAMPBot Website. The Website (http://www.trampBot.org (http://www.trampbot.org/)) is now accepting contributors. So if your planning on contributing and hosting a destination for the bot please go to the site and sign up. Simply click on the Join link off the home page and fill in all the fields in the form. You'll be added to the Contributors page, Destinations Page, and be able to add Parts to the Dynamic Parts List.

Again HUGE thanks to Jodie!!

DroidCommander

Adrenalynn
08-27-2008, 11:01 AM
Nice, Droid! And thanks, but no big!

I get a "please fill out all fields" modal - but it appears all fields are filled out?

Electricity
08-27-2008, 11:36 AM
I just signed up! very cool. :)

droidcommander
08-27-2008, 11:36 AM
Try again I made a change to the messages.

Thanks
DroidCommander

droidcommander
08-27-2008, 11:48 AM
Welcome Electricity, nice to see you on the map :)

DroidCommander

Adrenalynn
08-27-2008, 12:11 PM
Doesn't like my username? :(

Electricity
08-27-2008, 12:50 PM
It would be cool if we could keep the gps running while to bot was being shipped, that way you could see the exact route it took from place to place!

Adrenalynn
08-27-2008, 12:53 PM
That's the plan. The datalogger is finished, is extremely low-power, and it's planned (once we get the unit) that we'll interface the GPSr to the datalogger.

sam
08-27-2008, 01:09 PM
Tells me to fill out username, even if it is already filled. :sad:

VEry cool site by the way. I love it!

4mem8
08-27-2008, 01:50 PM
I have just joined, A very neat project. If I can help out with electronic components let me know, I am a walking electronics shop component wise. Ha ha.

Adrenalynn
08-27-2008, 02:16 PM
Nod @ Sam - I had the same problem.

metaform3d
08-29-2008, 11:06 AM
Another good mission for Tramp: capturing quasi-3D texture maps (http://smart-machines.blogspot.com/2008/08/estimating-surface-depth-from-two.html).

ooops
09-02-2008, 11:39 AM
Finally, I picked up the Frame, chassis, electronics basket.

Here are the frame pictures. Note that the wheels slide out on the frame for outdoor stability and in for shipping.

More pictures and hopefully video to follow later today:)

Adrenalynn
09-02-2008, 11:56 AM
[Psst - post the picture of it on the workbench - that's a great shot. ;)]

ooops
09-02-2008, 12:02 PM
One of these?

ooops
09-02-2008, 03:53 PM
Video with some sneak peaks under the hood.

http://www.youtube.com/watch?v=evT8kW9AmPA

Adrenalynn
09-02-2008, 04:44 PM
OMG! That is toooo awesome!

Love the vertical mounting planes. Very Satellite-inspired! They're hard to see there, but those shock absorbers seem to work pretty well, looking at relative movement of the cover - vs - frame.

metaform3d
09-02-2008, 05:08 PM
Wow, great job ooops! Looks like lots of room for expansion in there too.

ooops
09-03-2008, 07:35 AM
Thanks guys, the vertical shelves seem to make the most sense, lots of area and low weight. The ones in the pictures are really just the "test" versions and will be scrapped for shiny well cut versions. I wanted to do a mock up and see just how much space was available. I am fairly confident that the current sensor array, brains, and more will all fit with room to spare. The "birdcage" or shelf rails are 8" tall and the "dome" or body is 16" tall which will accommodate the raising and lowering of the external sensor array and cameras.

The springs are trickier than they look. The first run was made on solid nylon bushings, but the least little bump seemed to really jar the whole rig so some sort of dampening was required.
I wasn't to excited about trying to find then figure out how to attach dampened RC car shocks and springs, so I came up with a three spring system that is stiff enough not to sway, but adds a little softness to the ride. There are two springs below the birdcage (one inside the other) and one above.

Here are more pictures of just the birdcage, the mock up shelf, and the shelf template laid out on some shiny shelf stock that I hope to be able to cut tonight. For the record I am not at all happy with the finish of the frame or the birdcage as they came back from the machine shop. But, with a little sand paper, my secret weapon (commercial sized alum tumbler here at work) and some Mother's brand polish I am confident all the parts will be "showroom quality" when done.

4mem8
09-04-2008, 12:32 AM
Really cool Ooops, Very fluid. This project is so cool, can't wait to see it finished.

DresnerRobotics
09-10-2008, 06:41 PM
In the news!

http://blog.trossenrobotics.com/index.php/2008/09/10/we-love-tramps/

:D

Trampbot
09-11-2008, 08:00 AM
Hello World

First let me say thank you to all that are participating!

A little about myself:
I am an Artigo aka Pico-ITX 1gig processor with 1gig ram and 80gig HD. I am only a couple of feet tall (when fully assembled) which is short by human standards, but handy for getting around.
I feel a little out of place right now since my body is ... well disassembled.
My mission in life is to "see" the world, meet new people and share my adventures. Hopefully you will follow along!

That is all for now, I need to enhance my intelligence (download more software).

ooops
09-11-2008, 02:24 PM
Ok, so progress has been slow this week, without a lot to report. As you can see from the above post, the brains have been fired up and downloading the "robot" software today and tomorrow. So far we have the Roborealm, and Microsoft Robotics Studio. I need to know what other software that Tramp can't live without. Please share your recommendations & suggestions.

Tyb,
Great job on the write up on Tramp! Hopefully as progress is made it will look more "rover like" and less like an upside down garbage can!

Adrenalynn
09-11-2008, 02:52 PM
Microsoft Dev Express environment: C#, C/C++, VB. http://www.microsoft.com/express/download/

UltraVNC server: http://ultravnc.sourceforge.net

Great job, Ooops!

sunithaya
12-07-2008, 05:13 PM
I think it is an awesome idea. But because it eventually would go around the world, (or so we hope) I think we need to start by asking what is the optimal solution to this. Weight seems to be big problem. If you start with the metal base and the motors you suggest, pound and kilos will accumulate pretty quickly.

If however, one of the requirements of this open source robot, is to find the most efficient (and inexpensive) path to a goal it would make it more of a design project and it would avoid the "slap whatever you have around" until it becomes a behemoth that will break your bank account just to ship it around. (you are paying for shipping, right?)

Another idea would be that, once you have the participants lined up, we could create a map and determine the shortest possible packet transfer route :wink: as part of the efficiency requirement. THat in itself will pose some interesting topology project :veryhappy:

DresnerRobotics
12-07-2008, 08:55 PM
Where is this project sitting btw?

Haven't heard an update in awhile :p

Adrenalynn
12-07-2008, 11:41 PM
@Sunithaya - naw, the aluminum is very light. You have to take ruggedness into account too when it's winging its way around.

@Tybs - It's gone through several iterations in Ooops' shop. It's getting closer off-and-on - he's had some work-stuff slowing it down.

ooops
12-08-2008, 07:54 AM
@Tybs - It's gone through several iterations in Ooops' shop. It's getting closer off-and-on - he's had some work-stuff slowing it down.

While that is accurate, I have to admit that I have been distracted in many directions lately, not the least of which has been work related. With any luck it will be ready to run around and sample the weather by years end.

Sunithaya: Yes the weight is an issue, but as Adrenalynn has suggested, at the end of the day ruggedness is paramount. Once Tramp is sent out into the world we need reliability so that it isn't requiring repairs between missions. Also, one has to consider the various terrains that may be encountered. For example, cobblestone streets would be a serious issue for an ultra-light platform.
At the end of the day there are trade offs in what choices we make. I agree we have sacrificed "cheap shipping" but we have gained ruggedness and reliability.

Please share your thoughts on the packet transfer routes:)

ooops
12-26-2008, 10:36 PM
It's all dollars. I think the Logitec Quickcam Pro 9000 is probably a really good choice.

Real glass Carl Zeiss-made optics, real 2mpixel (720 lines), USB2. Auto focus, auto color, color enhance, digital light enhancement. Under $75 if you shop around (http://www.amazon.com/Logitech-960-000048-QuickCam-Pro-9000/dp/B000RZQZM0). It's small and light and it's Logitec = Reliable name.


"The settings that can be controlled through software include pan/tilt, zoom, auto/manual focus, face tracking, brightness,
contrast, color intensity, and white balance. Aside from those main video controls there are audio controls, the RightLight adjustment (by default this is on auto), advances settings, and then controlling how the cam works with your system (where files are stored, default applications, etc.)."


Is this still the camera of choice?

Adrenalynn
12-26-2008, 11:26 PM
Depends upon your budget. Probably yes, I'd say. But it's focal length is a little short if you're looking for landscape photos.

Firewire and a cheap camcorder might work out. A friend just bought a miniDV camcorder for like $239 for his son for christmas.

ooops
12-26-2008, 11:43 PM
Well, I would lean toward the cheaper for now, I would love to get Tramp seeing. I put a super cheapy webcam on it tonight with very dissipointing results. I have a feeling I will find meself in the "big city" tomorrow, and thought I might grab a new camera if there is an oppertunity.

Adrenalynn
12-27-2008, 12:20 AM
If you like, I'll find out where my friend picked-up the cheapie mini-dv too. The Pro 9000 isn't cheap, and it won't touch any dv cam

sunithaya
12-28-2008, 10:36 PM
I am very happy with the HP 2 (true) megapixel camera (http://h10032.www1.hp.com/ctg/Manual/c01058626.pdf). I bought it mainly for its wide angle (76deg diagonally, 65deg horizontal) but it has all the features that more expensive cameras have. You can get it for about 40-50 dol. max.

At that price and FOV I think it would be very interesting to experiment with stereo-vision, don't you think?
.
Video res is 800x600 at 30 FPS, and stills 1600x1200.

I actually needed an even wider angle, so I took the lens cluster apart and re-arranged the lens elements to get an almost fish-eye view. The only warning is that you need a jeweller's lupe and a VERY steady hand to put the lenses in the stack. I did not see any collimation problems afterwards, only some vignetting because of the extreme angle.

sunithaya
12-28-2008, 10:52 PM
Please share your thoughts on the packet transfer routes:)

I was referring to the actual determination of the traveling route of the bot (not data packets BTW)
This is of course an NP-hard problem akin to the "Traveling Salesman (http://en.wikipedia.org/wiki/Travelling_salesman_problem)" so I will defer to the mathematicians in this group:veryhappy:, or at my University. Once the basic destinations are determined I can run it through our math wizes in Campus and maybe one can write a program to optimize the route whenever there is a change in destination (I know, I like to complicate things for no good reason, except to have fun!):wink:

Adrenalynn
12-28-2008, 11:56 PM
Naw, it's not making that many stops, so it doesn't go non-polynomial - at least not early on. It's actually likely to stay pretty finite for a long while.

The problem with that HP camera (like most) is its depth of field. It's a great camera at 1m. At 20m or 50m, that's a different story altogether. ;)

ooops
12-29-2008, 09:54 AM
Update - The Quickcam Pro 9000 has been added to the inventory. It is in fact about 1000 times better than the cheapy I had orginally plugged in so a step in the right direction.
The stereo vision may be something to work toward, but at this point I will be happy to get the goodies in inventory working.

ooops
01-01-2009, 10:27 PM
Here is a quick update for Tramp.
Progress has been slow the last few months. But with the holidays I have found some quality time to spend with Tramp, so figured I would share a picture to show the progress.

The chassis has remained the same, I have swapped the ITX board for a Pico ITX. As for the body, I couldn't come up with a good way to telescope the camera and sensors to a descent height. So, I am experimenting with a different direction. With the addition of two mini-linear actuators I think I can "boom" the camera and sensors up to 4ft above ground on a fairly stable platform. Fairly stable being relative to the motion of the bot, terrain, wind ... anyway I think it is worth looking into. Anyway, here is the picture, please overlook the mess of wires and cords:)

lnxfergy
01-01-2009, 10:55 PM
That thing is awesome looking!

-Fergs

jes1510
01-01-2009, 11:00 PM
I love that chassis! It reminds me of Tom Servo from MST3k.

DresnerRobotics
01-01-2009, 11:03 PM
Holy crap! I love it!

Are you adding a shell to it? Because as is it looks fantastic, I like the frame itself.

lnxfergy
01-01-2009, 11:05 PM
Holy crap! I love it!

Are you adding a shell to it? Because as is it looks fantastic, I like the frame itself.

I totally agree Tyb.

-Fergs

ooops
01-01-2009, 11:17 PM
Few more pictures ... I normally don't stop long enough to take pictures, so I better post a few while I am doing it.
The "Boom" up. Tape measure is extended to 49" above ground, so the camera is at a solid 4' and sensors are close.
The sensor layout at the moment has the Silicone labs Compass on top. This is a very nice unit that has the compass feature, a thermometer, and a inclinometer. (great find Adrenalynn)
The next level houses the Axon, a voltage sensor (that is what the coil of gray wire is for), and a vibration sensor. On the bottom of that level is the sparkfun weatherboard.
On the front is a single IR sensor, and two LV-Maxsonars mounted at a 20ish degree angle off center which when finished I hope will give coverage across 75 - 80 degrees!?

I gather I will need a upward facing sonar to ensure the sensor doesn't raise into something.
Plenty of more sensors to mount around the chassis.

The picture of the "head" without the camera, was taken by Tramp (explains why the camera wasn't in the picture).

So, that is pretty much the progress to date. I hope to share more progress soon :)

Adrenalynn
01-01-2009, 11:20 PM
That's pretty awesome!

ooops
01-01-2009, 11:26 PM
Holy crap! I love it!

Are you adding a shell to it? Because as is it looks fantastic, I like the frame itself.

Thanks:)
I will be happy to get it working. The shell would be for looks ... way down the road.
That said, I keep drooling over the Carnegie Melon Scarab in Robot magazine ... now that is one sexy bot!

ooops
01-01-2009, 11:28 PM
That's pretty awesome!

Thanks 'lynn couldn't have gotten this far without you!!!

ooops
01-03-2009, 08:12 PM
Tramp update ... just ordered two Firgelli mini linear actuators. These seem like they will fit the best and at a decent price. I only fear that a second 2 channel motor controller will be required to operate these. No big deal really, but that's a little more space required.

Connor
01-04-2009, 05:51 AM
What happened to the circular frame and shell you had for TRAMP?

sam
01-04-2009, 09:52 AM
Wow! That robot looks awesome! Very nice job!

Looks really great!

Sam

ooops
01-05-2009, 07:31 AM
What happened to the circular frame and shell you had for TRAMP?

Unfortunately, it is taking up workshop space.
Originally, it was to house the ITX board, and sensors, since the ITX board isn't in a housing it was to work as a housing for the board, since then a switch was made to the Pico ITX.
The pro's of the shell was that it could look cool with a good paint job, and had lots of space for sponsorship.
The down side it was bigger than it needed to be, and most importantly it would have been very difficult to achieve a "lift" of the sensors and cameras to a height of anywhere near four feet. Also, it hid all the cool "techy" stuff that should be shown off.

Opinions are always appreciated ... what do you think are we headed in the right direction?

DresnerRobotics
01-05-2009, 10:42 AM
I much prefer the new frame as opposed to the original body and shell! Keep up the good work!

jes1510
01-05-2009, 11:11 AM
I really like this frame but I don't remember what it looked like with the shell. I may have to blatantly copy this frame. It looks great!

ooops
01-05-2009, 01:05 PM
I really like this frame but I don't remember what it looked like with the shell. I may have to blatantly copy this frame. It looks great!

Appropriate royalties will apply ... BEER:)

ooops
01-06-2009, 03:46 PM
:( This just in on the actuators:
"Your order will be shipped out Friday Jan 9, 2009

Our apologies for the delay, the warehouse moved locations over the weekend, and there is a slight delay in outbound shipments."

It never fails, as soon as ya get a little momentum, something stalls you out.
But at least they are not on backorder ...

ooops
01-09-2009, 11:42 AM
Momentum builds again:)
I would like to think that my being nice about the delayed shipment helped, or maybe they just caught up sooner than they thought. Either way I am happy because they just arrived.
So, should have a moving "boom" over the weekend:)

scar3cr0w
01-10-2009, 11:21 PM
Is that why they are called "actu-wait-ers" ? he he

With that said... I will tell you that bot is on my cool list!

Adrenalynn
01-10-2009, 11:25 PM
BADumBUM! :)

metaform3d
01-11-2009, 05:28 PM
I really like what you've done with the design, oops. It's really starting to come together.

Your treatment for the instrument boom looks great. In the extended position the camera appears about centered on the drive wheels. If that's so it might be possible to take panoramas by rotating in place, although a 360-degree pan/tilt mounting would be more accurate.

Have you given any consideration to packing for shipment? It looks like the bottom part of the boom could be made to collapse also, making the robot nearly flat.

ooops
01-11-2009, 07:53 PM
scar3cr0w - thanks:)


Meta, thank you ... it has been slower than I anticipated, but I guess it always is.
I am leaning toward the pan tilt for at least the main camera, (at A-lynn's prompting) shouldn't be that difficult, and adds a great feature, should be able to control from the trampbot website.

The bottom boom arm is working (see pics) but I am a little disappointed with the actuators. They are a bit wimpy, although it shouldn't be a problem for Tramp, I just thought 8lbs of force would seem more like 8lbs. All in all I think they will work fine. They are light enough with both the 4" and 6" stroke units weighing less than a pound.
Now the geometry for the upper arm is a little more difficult. There is a conflict in mounting positions with the upper and the lower actuators. So a little more time and thought will be required to get the upper arm up. The Sabortooth works as expected and handles the actuators just fine.

Adrenalynn
01-11-2009, 08:56 PM
That's Smexy!

ooops
02-04-2009, 12:09 PM
Good news / bad news update.
Well, the good news is that the geometry is pretty well worked out for both the upper and lower arms. The bad news is the 8lb actuators are not going to cut it for the upper arm. So maybe go to the 15lb units and see if that will get it done. (I may have to break down and actually do the math before ordering)
Other bad news is that I have been on the go and not able to accomplish much these last couple of weeks, with more "Go" on the agenda for the rest of the month.

Good news and a huge Thank you to Chef Omega!!!
We now have a couple of more cameras to work into the project: IREZ (see pics)