All terrain mobile platform

As a small group of hobbiest we want to build an outdoor mobile platform with the specs listed below.
Can you help us by sharing your experince to guide us on any of the topics listed below and/or any additional topic that is not listed?
We beleive your contribution will be appreciated not only by our group but all other hobbiests and robotics gurus
[Main computer] A notebook PC with USB ports (1GHz, 256MB RAM, 20GB HDD)
[Batteries] Operates with 12v or 24v (car) batteries
[Load capacity] Carry upto 50kg (or more)load
[Terrain travesibility] Drive over 20cm high obstacle (rocks, kerb, etc)
[Speed] Speed upto 0-20km/hour (or higher) forward and reverse
[Localization] Able to localize itself within 10cm accuracy (GPS + what else?)
[Heading] Able to detect heading angle in 5degree accuracy (or better)
[Comunication] Communicate with a base station at a distance 5km (or more)
[Video transmission] Able to send video to 100-200m (or more)
[Motor] DC motor sources: http://www.oatleyelectronics.com/bcorner.html [Motor Control] ?? [Your additional topic] ??

snipped-for-privacy@yahoo.com.au wrote:
  Quoted Text Here  
Wow! These look really nice (from the above page):
250W DC GEARED MOTOR
US dollar price is something like $55, which is pretty darned good. Shipping would be a PITA because of weight.
Only thing to add, based on selection of these motors, is to get four of them, and connect them via short lengths of roller chain to your 10" or larger pneumatic tires. Be sure not to attach the motors directly to the wheel shaft, yada yada.
Not too thrilled about using a notebook PC with a spinning hard drive on a vehicle that s used for outdoor 4WD activities, unless you get a ruggedized high-G drive (EXPENSIVE). I'd recommend a lower-tech magnetic drive, and/or some Flash drives for storage. Stay away from any system where you'll get an HD crash when you go over a big rock at 5 MPH.
-- Gordon Author: Robot Builder's Bonanza Budget Robotics: http://www.budgetrobotics.com

Definitely...
You may want to consider a mini-ITX system with a flash drive like the ones they sell at www.mini-box.com.
I think these are much better suited to mobile robotics than a laptop.
BRW
On Thu, 07 Oct 2004 16:26:29 -0700, Gordon McComb
  Quoted Text Here  

Well I use a Via Mini-ITX (http://store.ituner.com/ituner/viamo.html ) with a serial port controlling a ISOPOD (www.newmicros.com) controlling a couple of 100 amp motor controllers like this old test bench photo shows: http://tinypic.com/blhsiThe motor controlling system is where it is all at, you cannot skimp or cheat on that. You need good motor controllers, that's where all your money goes. The Mini-ITX board uses flash drives so that bumps and vibration aren't a problem. The motor conteoller I show is one of my own designs, because it can be very hazardous, I haven't made any of these big motor controller designs public yet. Now you also need some I/O for the Mini-ITX so a USB to TTL I/O adapter works nice, there are a number to choose from like this one http://www.dontronics.com/u401.html or like this one http://www.dontronics.com/giga.html#io24. Of course since this is outside a GPS unit with a serial interface is a must have item too. There are a lot of these to choose from too. Don't forget old fashioned bumper switches as you don't want it running over anyone it doesn't see with the other sensors. Most important is a reliable kill switch, that you can easily reach in a full run while your chasing it across the park :)
<Bennet Williams> wrote in message wrote:
  Quoted Text Here  

Thank you all for sharing you experiences with us.
Having a good, high power motor controller is very important. Your 100 amp motor controller at http://tinypic.com/blhsiseem to be great. Altough you indicated your reason why you do not its design public, may I ask if you can share its schematics with us so that we can lear from your experience. (I promise we will not sue you when we blown-up our circuits:))
How do you count quadrature encoder signals ? Do you have quadrature encoder counter(s) on your board or is it ISOPOD counting the pulses? Do you have PID controller code running on ISOPOD?
  Quoted Text Here  
, , ,

The ISOPOD can decode the signals output by Quadrature Decoders automatically (it is built in). I just feed the outputs from the quad decoders right on into the ISOPOD. Then the MCU just keeps track of the distances for me. The ISOPOD uses a DSP processor that is pretty much designed for advanced motion control.
I don't know about letting people know what I did.for the motor controllers. Fortunately there are several sources, so I don't feel there is a need for me to also publish information about them. Basically, it is dangerous, for one example, a short circuit can cause the batteries to explode. It isn't the motor controller that is the weak point. Testing and debugging has to be done carefully with a good current limited power supply, as the transistors basically measure on resistances less than the solder joints themselves. Thus while your trying to figure out why it is not working the battery packs explode on you. Yes there are fuses, but at 10:00PM at night and you've blown all your fuses, what do you do? In the photo I took, you'll note the current limited power supply, and the little tiny wires used during the initial testing phases. So if there was a problem, it wouldn't blow up anything easily. Fortunately, it worked Ok, although I needed to allow more room for bigger capacitors on the PCB. :) Nicads, Lithium Polymer or lead acid batteriries all become very dangerous when shorted out. One of the things is the motor and battery power wires need to be able to handle the high current levels. The MOSFETS I used in the photo don't start getting warm until you exceed 380 amps, the wires I used can handle 300 amps no problem. Then while your standing there looking at it wondering why it don't work, kaboom goes the battery packs. The other thing is a big robot is heavy and dangerous, you just have to be careful all the time. A safety checklist you go through before you do anything is important. I have benn using IRF IR2104 half bridge controllers with IRF MOSFETS and I like the way they work. I like using locked anti-phase PWM nowadays, so they fit in nicely for that. The circuit I used is pretty much the same as what IRF shows in the data sheets. Others like using the OSMC project http://www.robot-power.com motor controller system. When I first started builiding my motor controllers a few years ago, you could not get the IC the OSMC project people used so I used the IRF2104 chips instead. Chuck McMannis also has a nice article here about his adventures http://www.mcmanis.com/chuck/robotics/projects/esc2/speed2c.html with high power motor controllers. The IRF2104's have a disadvantage in use though, since the PWM signal also charges the high side charge pumps, you can't go much over 90-95 percent duty cycle or the high side transistors won't be driven properly. But for me, full power isn't a problem as I haven't had anything need more than half to 2/3's full power yet. I can't run 20mph+ so my robots don't go much faster than 5mph otherwise I can't catch them. I get nightmare visions of my son and I chasing a beserk robot around the park trying to reach the emergency shutoff as the robot dodges trees and I don't. :) Of course if you read this far, my Park Robot Number Three is getting a suspension system. http://tinypic.com/blulh
<Roberto> wrote in message >> Well I use a Via Mini-ITX (http://store.ituner.com/ituner/viamo.html )
  Quoted Text Here  

I forgot You can goto www.newmicros.com they have code examples of doing PWM, PID, Quadrature Decoding, etc for the ISOPOD.
  Quoted Text Here  

THANK YOU for the information. If I'm not mistaken you seem to used 8 MOSFET in http://tinypic.com/blhsi. Can I ask which particular MOSFETs (IRF ???) you have to achive 100Amp? Do you have some other URLs to see more photos of your platform?
  Quoted Text Here  

Actually I've used several different MOSFETS the PCB layout works with most any TO220 cased n-channel MOSFETS. The IRF3803 is really neat RDS .006ohms, 30v max, 140 amps continuous, 470 amps pulsed max. You usually don't even need heatsinks. Disadvantage is the 30v max rating though. The IRF540 classic is still really great too, RDS .077ohms, 100v, 28amps, 100v , 110 amps pulsed max inexpensive and makes for great general purpose small motor drivers (small motor is relative here they are still big). The IRF1010 is still a really excellent motor driver with 60v max, RDC .012 ohms, 84amps max, 330 amps pulsed max. Using thr 96v to 130vDC motors I used the IR3415 high voltage MOSFETS to good effect. Now this is extremely dangerous as high voltage can kill you.
<Roberto> wrote in message >> The ISOPOD can decode the signals output by Quadrature Decoders
  Quoted Text Here  

THANK YOU for the information. If I'm not mistaken you seem to used 8 MOSFET in http://tinypic.com/blhsi. Can I ask which particular MOSFETs (IRF ???) you have to achive 100Amp? Do you have some other URLs to see more photos of your platform?
  Quoted Text Here  

Gordon McComb wrote:
  Quoted Text Here  
With a large vehicle like this one sounds you can afford the space needed to make a cradle that will isolate the computer from the worst of the shocks.
Solid state memory like flash cards are fine except they are fairly expensive and have limited lifetime (limited writes anyway). Magnetic tape is slow. Ideal solution would be system with a large amount of conventional ram (2gig would be good) where you can have a ram disk where you keep stuff that is used a lot. Ideally you could have a conventioanl disk drive that is used to boot the system (while stationary) then have it spin the disk down and park the heads and run off of the ram drive when it starts to move. Hard drives are bloody tough when parked; I think they can easily take something 300g shocks when parked - this is how those laptop drives can "brace" for impact and survive if you drop the laptop! Actually, sod all this and just get one of those laptops :-)

Matt Dibb wrote:
  Quoted Text Here  
you are referring to ibm thinkpads with the accelerometer? i believe the way they work is by detecting that they are falling (ie. no acceleration) and park the disk heads before they hit the ground. that probably wouldn't help in this case because in the event of going over a bump there would be no such warning.
-eck.

Edward C. Kern wrote:
  Quoted Text Here  
Well, I dont think they detect no acceleration as a prompt for parking the heads otherwise they would be parking the heads when they are sitting on a desk etc. Probably when they detect acceleration from 0 to 9.82mss or whatever gravity is then park the heads. Chances are if its accelerated to that its gravity doing it ie its being dropped.
Might not just be gravity's acceleration anyway, jsut something over a certain rate, probably acceleration over a level that is likely to cause damage, in which case it will park the heads whatever. On vehicle though it might not be fast enough to cope with the changes in acceleration - dropping from 1m up to floor gives it a chance to park the heads..

Matt Dibb wrote:
  Quoted Text Here  
You do understand that an object on a table experiences around 9.8 m/s2 just from gravity?
When the acceleration goes lower than that, it is falling. -- D. Jay Newman

D. Jay Newman wrote:
  Quoted Text Here  
Sorry just got back from the pub when posted, tend to revert back to some beer-fuelled "common sense" at that point. I guess I decided that no acceleration meant "not moving" having reviewed my post! :)

Matt Dibb wrote:
  Quoted Text Here  
I'm also sorry for being a bit snippy. I have an illness in my family which is taking up most of my energy. :( -- D. Jay Newman

This is a solution looking for a problem. Few robotic applications require the memory space of even a few gig. Windows is ill-suited as a real time operating system, so its disk footprint is not required. At most a hard drive could be connected for programming, and any Flash device should be good for 1,000 writes. How many times would the thing need to be re-programmed? State information from boot-to-boot could fit on a high-density floppy.
On the surface it sounds good to provide as much memory as they sell on the average laptop, but it's a bad habit to get into. Robots running around loose need to act fast. The larger the memory requirements the slower the thing is. Keep the code tight. Look for ways to compress databases. The less memory needed the faster the processing.
Consider the "hunker down" scenario you've outlined using an accelerometer. Good idea except in a moving vehicle over rough terrain the hard drive would be constantly parking. Not too useful, IMO.
-- Gordon Author: Robot Builder's Bonanza Budget Robotics: http://www.budgetrobotics.com
Matt Dibb wrote:
  Quoted Text Here  

Gordon McComb wrote:
  Quoted Text Here  
Well the original poster didn't tell us what was required. One would think that if they are going to be putting a laptop on the robot they would want to make use of it for something more than just some basic control routines. I was at one stage planning a similar off road robot thing but more as a poor-man's mars rover. One idea was the robot would trundle around on its own and occasionally phone home with the data and wait for instructions or just carry on regardless; that sort of situation requires a lot of storage. Also the software and memory required for things like fairly high-resolution image processing for navigation purposes etc are not insignificant.
Anyway the idea wasn't to have some huge bloated amount of software filling 2 gig of memory, you would partition it off and use it as a hard drive, so you keep 128meg or something for OS and software, then use the rest as a storage place for data etc (the ram disk). The idea of this is you can read a load of files from a mechanical drive into the ram disk and use it as a general storage area. This means that you can spin the drive down after the machine has booted up (or dont even leave it connected or on the robot at all) then have it running off of solid state memory for the duration. The significance is that computer memory is cheaper and longer-lived than something like flash memory drives.
You could use some flavour of linux etc, or even something like QNX RTOS if you are feeling exotic, for the operating system and par it down quite nicely as far as memory foot-print is concerned anyway.

Matt Dibb wrote:
  Quoted Text Here  
But this is the part that doesn't hold up to scrutiny. You are thinking of a hard drive intended for a static, safe application. A ruggedized hard drive, like those used by the military, cost many thousands of dollars. They are made just for this type of thing, and standard commercial drive won't cut it. Flash memory, even with a limit of about 1,000 write cycles, will still outlast the average hard drive in an abusive environment.
Two gigs of Flash would cost only a little more than the typical non-ruggedized hard drive. It's used for storing the OS and programming. RAM and EEPROM are used for program space and data, as they are in any typical microcontroller.
You must go solid state with a heavy robotic vehicle used in the "outback" (my estimate is that with a 25-35 horsepower gas engine this vehicle would weigh in the neighborhood of 400-500 pounds, minus its 100+ pound payload). Failure-mode risks are just too great otherwise.
-- Gordon Author: Robot Builder's Bonanza Budget Robotics: http://www.budgetrobotics.com

Gordon McComb wrote:
  Quoted Text Here  
No I am talking about computer memory as in RAM, not a hard drive. You can get a gig of pretty standard pc2700 DDR ram for about £90 inc vat and postage. You can run a ram drive in that and it will be long lived, robust, extremely fast compared to flash and reasonably cheap. A gig of flash memory is about £60 for comparison. The idea is you use a conventional hard drive to load data into it at boot time then either totally remove the drive or spin it down and park the heads before you start.
I've no idea about the power consumption issues but if you are going to be running a pc on your robot then you probably already have more than enough.