Tablet PC as brain?

Hi, I have been looking for a good platform for an autonomous robot project, and it requires a fairly full PC configuration where I am
able to run Java and wlan access. I have basically looked at the Mini-ITX systems (and now the Nano-ITX systems look promising), and a 1 GHz system cost around $200, excluding the necessary battery power needed, wlan and disk drive. I dont know how long you would be able to power one of these on e.g. a 3800 mAh lithium cellpack. Anyone have experience with this?
Now I have also been looking at more expensive solutions where a full blown TabletPC would be used, enabling a touch screen interface in addition to the custom vision/speech recognition system. A model like the Compaq TC1000 with a 1 GHz Transmeta CPU sells for under $1000 on eBay now (probably refurbished models) and it seems fairly well spec'ed, wlan, cfport, 2 usb2.0, 1 firewire, sound line-out, mic-in, 768 MB ram, 30 GB hd and detachable keyboard + more. It sounds like a pretty good package with enough IO to control the rest of the robot (USB hubs can add to this). The whole thing could be embedded vertically in the robots torso. At 1.36 kg including a Lithium battery for 3 hour operation it sounds like a fairly lightweight "brain" (considering the screen would probably be off a lot of the time).
Anyone tried this?
Best
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
This is the easiest way to go, if cost is not an issue. You don't have to solve battery and power supply issues and you get compact, decent performance PC off the shelf. Consider Pentium M based tablets (like Acer TravelMate C110/C111 or Compaq TC1100), if you can spend a few hundred dollars more. You will get several times better performance than TC1000, especially if you want to do some image processing.
Paul.
snipped-for-privacy@online.no (Jeceel) wrote in message

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Try a Fujitsu Stylistic 1000. It's only a 486 sx @ 100mHz, but not very heavy or large and a great battery. I used to use mine for my classes - it was my grade book/lesson planner and held Power Point demos for an LCD panel. It got between 4 and 6 hours on it's battery, depending on settings and use. I bought three on ebay for $75, a great deal.
chris in napa
Paul Jurczak wrote:

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

A Tablet PC is not a touch screen. It has an electrostatic display that pulses an energy field across the entire screen very rapidly that energizes a small capacitor in the stylus (the pen), when the pulse goes away, the cap discharges powering a small circuit of some sort which emits some kind of 'becon' that the PC uses to know where the been is. Touching the screen with your hand has no effect on anything. Its like this so you can rest your hand on it and write without activating anything. The stylus will move the cursor even though it is not touching the screen.
Now if you dont mind using the stylus to touch things on the screen, then this will do the trick. Otherwise, I suggest sticking to a miniITX and a stand alone touch screen LCD.
-C http://hossweb.com
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Ok, I didnt know that. In that case the touch screen would be of little use as I would have made a user interface with large buttons on it that a user should be able to press with his fingers.
Actually, the screen is just a bonus, it wouldnt be necessary for the project I was thinking of. The computing power is what interested me first, and especially the battery life. These things claim to run for 3-4 hours and thats including the powering of the LCD screen.
The nice thing about these PC's is that I do not have to worry about the power and all the electronics for regulating power and charging is already there, and it would be pretty simple to create a docking station for the robot to park for more power. I read that people say that you should avoid Lithium-Ion batteries for robot projects as it is to dangerous if you do not know what you are doing, charging them in a safe place just in case they explode. Well, I do not want to take these risks as I am more of a software guy than hardware, so I was looking for complete powering solutions where I didnt have to worry about these issues. NiMH batteries will add so much to the weight that I am afraid it would bog the robot down (servos and such would then have to be equally powerful).
Does anybody have any idea how it is to power the Mini-ITX systems with a Lithium-Ion battery of around 3800 mAh as most of these portable pc's use? They have a calculator at www.mini-box.com that show me around 36watts drain when running a configuration with two USB devices and one 2.5" HD in playing DVD mode. I would assume my robot would be running 100% CPU almost all the time, busy with its vision processing all the time. How do I go about calculating how long a 3800 mAh battery would last on this? (I am sorry if I am reapeating myself, but maybe someone have a quick answer to this).
Thank you.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Ok, I have done some more researching and a typical Mini-ITX system would run for about 12 hours on a 7 Amp SLA battery which would give it a mAh/kg ratio of 2641 mAh/kg. These numbers are based on BB's BP7-12 battery. Its still a heavy battery for a robot 2.65 kg although these batteries are cheap. I want to reduce the weight as much as possible as there will be times the robot needs to lift itself up to e.g. climb stairs or similar obstacles in a home (typically this would be an issue going through doors too).
Some more research I found 12 volt 3300 mah NiMH at a robot webstore that has weight of 0,62 kg resulting in a ratio of 5322 mAh/kg. So a doubling in power per kg. Two of these would give me 6,6 Amps and around 10-12 hours operating time on the Mini-ITX with some USB devices and a 2.5" harddrive. Of course, the servos and motors would also draw some power from this probably halving this again I would expect. Still around 5 hours of operating time on a 1 kg NiMH battery system seems acceptable to me. If the robot has the torque I could always add another pack and maybe get a full day of operating time.
Now the next challenge is where to find a good system where the robot could dock and charge the NiMH battery when it is close to depleted. This would require my robot to measure when the voltage is dropping? Anyone have any information on this? Something that can interface simply with a digital input on a controller. It would also be nice if the robot could draw its power from the "docking bay" charger while it fills up its battery so that it is still operating. Is this difficult to achieve including the safety needed so I dont fry the thing?
Any help appreciated. Thanks!
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.