water finder in dessert....

[rickyrt1]

Hi,

This is the first time i've posted.....

I'm trying to design an autonomous device which can search for water in the dessert. My problem lies not with the sensors or motor control etc but with the choice of control circuits for the device.

I have three main objects in the design....

1. Autonomous vehicles themselves, battery powered with countless onboard sensors, GPS and RF transmitter. (which do the surveying itself)

2. Intermediate unit which is self powered and placed near the unit to communicate with the autonomous vehicle via rf and serve as a recharging station when necessary, max range 2 miles.

3. Main control centre this is to recieve data from the stationary intermediate units over long range RF.

There will be about 10 such vehicles with 10 associated stationary intermediate units over a 100 sq mile area left for days at a time, recharging as and when necessary..

My problem lies with how to control these vehicles, whether i build a control circuit for each individual vehicle or use the dedicated PC and outsource the processing over the RF link, sending back specific instructions such as go forwards...

The vehicles themselves will be controlled by GPS interfacing with maps as well as onboard sensors, i envisage using dedicated PIC's to control objects such as the sensors, motors, servo's thus restricting inputs into a processing unit or transmitted over RF.

The vehicle will also include a low res camera in case the vehicle encouters an unexpected problem which the device is not programmed to recognise and a human needs to view it from the central control room.

I'm not sure whether to control the units by outsourcing the majority of the processing to the control PC and then telling the device what to do over the RF link. Interfacing with the numerous PIC's. Although to do this would require uninterrupted signal between the two and i can't guarantee this. Which means the device would be stranded if signal is blocked by the terrain for whatever reason, retrieving the vehicle is not acceptable.

My two options as far as i can see of having the control units onboard the vehicles is to use PIC's or an embedded PC unit. So far i think the several PIC's can do most of the functions i need although they i envisage they can't handle the low res camera and using the GPS system to interface with a map to control the direction. This is what i need to know and whether i have no choice but base my design on an embedded PC unit capable of doing everything onboard the vehicle, interfacing with the PIC's then only sending what's needed to be sent to the control room...

Anything anybody can add to this i would really appreciate...cost is a major issue to me, outsourcing everything possible to the main control PC would save me a lot of money. Or the possibility of using numerous PIC's interfaced with each other over a dedicated PC unit onboard would swap my dollars for development time. I really don't know which way to go with this one, anything anyone can give to this idea would be gratefully appreciated, i feel like im banging my head against the wall everytime i try to rationalise the problem. Knowing it's only possible to do this with en embedded PC would be great so i can start development and ignore the cheaper alternatives....

kind regards,

rick

 

[audioguru]

You are a 1-man NASA.
Good luck.

 

[The Mad Professor]

I would concentrate on the autonomous vehicle , perhaps looking at the Mars rover might help here.

Scrap the intermediate stage entirely by improving the rover design and elevating the control station [ or at least use a higher antenna]

If the rovers can talk to each other then data can be daisy-chained, in effect each one becomes an intermediate for the next.

PS:You fail to mention the operating enviroment , this is crucial information.

 

[Tarsil]

Hi,

This is the first time i've posted.....

I'm trying to design an autonomous device which can search for water in the dessert. My problem lies not with the sensors or motor control etc but with the choice of control circuits for the device.
rick

Cool Ideea but it will cost a lot of money. Something like....the budget for a small NASA op. How about 5-6 team's equiped whit Land Rovers, Food, WHATER (Lot's), Map and GPS recv?
BTW....how u'l detect whater?

 

[Oznog]

He already mentioned the environment- it's in the dessert.

**broken link removed**

I'm not sure how you could decide the water sensor is not a problem. This is quite a difficult problem!

There's a lot to look at in terms of autonomous vehicles. In an ideal more or less flat terrain, this isn't a big issue. Just a GPS and compass sensor. But rocks, embankments, trees, ruts, shrubbery, etc in 95% of the real world present far more problems. People have accepted "machine vision" as being essential to most real-world applications. Cameras, radar, sonar have all been used, with varying degrees of success.

 

[audioguru]

Don't some kinds of animals live in the dessert, oops I mean desert?
They could lead you to water.

 

[HarveyH42]

Wouldn't finding plants and trees indicate a source of water? Water also runs downhill, so at the base of those ridges could be a few pools of water. Maybe some low lying areas as well. Before spending a couple hundred thousand on high-tech toys, perhaps a little low-tech research...

You will need to build your rovers to handle rough terrain, because that is where you'll need to look, if you want to find a usable water source. You do realize that if you find water one day, it might be gone in a day or two.

Didn't catch which desert you are looking to survey. I think you can find maps for most deserts on Earth or even Mars. Most of the reliable water sources will be shown.

 

[audioguru]

I think the temperature in deserts changes a lot. Maybe enough for dew to form puddles at night.

 

[bigkim100]

I found water in my dessert yesterday....so I sent it back with the waitress.

Man...if your going to embark on a multi-million dollar project...you better start by looking a bit more professional in your presenation.
I would think that the first thing you better seriously look into...is a good spell-check software.

 

[Oznog]

It's kind of hard to figure how you're supposed to find water. A vehicle can only manage a very slow speed on typically tough terrain. Aerial surveys would make more sense since they cover thousands (probably millions) of times the area. Maybe you've got like a ground-penetrating radar which might need a ground-contacting vehicle, but it's just very unlikely to run across a water source randomly crawling around at 2 mph. If there's one source of water in a 10 sq mile area, it might take tens of thousands of vehicle-hours to find it if it can only see a few feet to either side.

 

[Hero999]

I didn't think radar was any good at penitrating the ground. I always thought sonar would do a better job. Anyway this looks like a very big project to take on, in fact it's many projects rolled into one. Good luck.

 

[Nigel Goodwin]

Probably the best method is the oldest, just use dowsing rods - no idea how it works?, but it certainly does - even water companies use them to find where to dig the road up.

 

[Hero999]

I did think of that, but then I thought no that won't work, it's nothing but folk law.

 

[Nigel Goodwin]

I did think of that, but then I thought no that won't work, it's nothing but folk law.

Don't you have mains water where you live then?, check your local water board - they use dowsing on a day to day basis - and not by people brought in, by the engineers who drive about in the vans!.

 

[Hero999]

That doesn't mean they were fully trained engineers.

I'm not convinced, read the good old Wikipedia article, https://en.wikipedia.org/wiki/Dowsing, make up your own mind.

 

[kchriste]

I'm not convinced, read the good old Wikipedia article

Dowsing works because no matter where you dig on this planet, you will hit water eventually.
I knew a dowser that claimed that he could detect anything, including gold, cancer, etc with a dowsing rod. He never was willing to lead me to the pot at the end of the rainbow. Always had some sort of mystical excuse.

 

[HarveyH42]

Well, you have to expect a few frauds in ever profession, doesn't mean it's without merit. I don't know if it works or not from personal experience, but it's been around for a very long time, with tons of sucess stories.

There are a lot of plants, bushes and trees that you only need to cut off a small piece, shove it in the ground, and it grows. Wouldn't the rooting end of the dowsing rod be in need of water to survive? Plants do seek water. I've got a vine that keeps growing toward my outside AC unit (condensation spills out next to it). It's a nice flowering vine, don't want to kill it off, but I have to catch the runners before the tangle in the AC. The main vine is at least 20 feet away from the AC unit.

Kind of guessing the origional thread is dead. Was kind of hoping to hear a few more details on how the rover would handle the rough terain, and the H2O sensors he'd be using.

 

[Hero999]

Your vine can obviously detect water vapour by monitoring the humidity, there's nothing unusual about that. Many plants can monitor humidity and there are electrical sensors that can too. What is much harder to do is detect a water source though a thick layer or earth which in itself probably contains some moisture. The only practical way I know of doing this is to use sonar or better still drill bore holes.

 

[audioguru]

Why look for water in a desert? That is the last place to find water.
Here in Canada there are fresh water lakes all over the place. We don't need to look for water, it is everywhere. I'll mail some to you.

 

[HarveyH42]

Fresh water in Canada? I'd have that all the pollution from America was being blown up there. Kind of miss the fresh natural water where I grew up, Florida water is really foul, and people actually swim in that sewer water. I spent half my life in Oregon, grew up on the side of a mountain. The southeastern part of the state is desert, which is where I learned of few survival skills. Guess that is why this land roving water project was a little odd to me. Its a lot of rocks and hills, much is rough on foot, vehicles even horses would be a bad idea.