Solar robot question

[GrimBot]

Hello, I don't know if this is covered elsewhere but I was wondering what method I would use to program a robot to work just off of solar power, no batteries.

How would I program a robot to keep track of what it is in the middle of doing so that the next time the sun comes up, it can resume its tasks? I know solar isn't very powerful for this stuff but I'm thinking that a bunch of little robots could do the job.


Ok you pulled my arm and want the "grand scheme" lol.     I want to convert construction vehicle toys into robots that can do some excavating, like an exo-skeleton that I install onto the toy.. dig dig dig..  dump dump dump.

Oooo.. side thought..  another robot hooked up to a solar battery bank that does nothing but shoots a laser beam to the photvoltic cells of the robots, this way the only thing that needs a battery is the laser emitter.

[Schlayer]

I'm not a prolific programmer and as such I cannot possibly definitively say whether or not that last art is doable, but in terms of purely hardware limitations but probably is. As for the rest of your question, all you need to do is a little code magic!
Outdoor solar lights often include a little photovoltaic cell (the solar cell) with a photo-resistor or photo-transistor (the light sensor) in the middle. This setup provides two key things: 1) The light knows when it is dark and turns on only then. 2) When the light is not on, it is always due to the fact that there is enough ambient light for the battery to charge.
So, you have a micro controller and you programmed in some digging routines for one specific robot (we'll use a single one for this example since it's literally repeating the same thing for each). You set it up such that your robot will always be digging or, realistically, until the battery drains. Your battery is hooked up to a little solar cell/panel and to the controller so the electricity from the cell ONLY feeds into the battery and the battery ONLY feeds into the controller; the other servos, motors and whatnot will be powered from the controller since we're talking smallish scale toys without high current or voltage requirements.
The hard part: electronics theory.
Photo-resistors change their electrical resistance in response to ambient light levels. A photo-resistor starts out with high resistance in the dark and its resistance drops as the light level increases. Photo transistors also respond to light levels, but for your controller it's easier to deal with simple electrical resistance changes since Ohm's Law (V = I*R) since your controller should be able to measure precise voltage differences through the photo-resistor. Whatever type you buy, you'll need to use the controller to find some baselines for light levels affecting the resistance. What light level is low enough that your solar cell no longer generates any current? This can be considered your 'daybreak'. What light level is what you want to call 'night' and start your digging process? What light level will you call 'day' when you stop digging? All these will need to be experimentally determined and then used as arbitrary resistance values for your code.

If this seemed like a super dense block of nonsense, the key info is that you have to make your robots act like outdoor lights and respond to variances in light levels to determine when to be digging or recharging. If you wanna get more complicated, you might be able to figure out a way to check your matter voltage using the controller and set the robots to stop digging whenever they go below a certain voltage/get too discharged. This would be a nice feature in any case, but if you are using Lithium batteries it would be essential as they can get damaged beyond usability from being discharged too much, and if you try to charge a damaged Lithium battery it may melt or explode. This is not very good for flesh, and certainly not for your robots.
I strongly recommend Arduino for you controller boards as they use an easy language to program, run off any computer OS, and are pretty cheap and reliable.

Last thought: if you were to implement this whole day/night cycle idea in your robots, using the laser or other light source aimed at the robots would be pointless since it would trip the photo-resistors and make them all turn off again. Much simper than aiming a laser is just bathing the whole area with flood lights, though it would look far less cool. The you would't have to worry about ever turning the robots off so long as the lights can keep powering them!

If you have any questions, feel free to message me I'll probably get back to you within a few hours.

[GrimBot]

Thanks, I have thought about that "floodlight" instead of using a laser but I keep remind myself that I'm trying to be efficient with it so the power I'm using to light all the areas that arn't doing the robot any good would just be an energy leak, I was just trying to use a laser pointer as a wireless wire, haha.

I do like what you said about choosing a work time using outdoor lights, I haven't thought about that. I'm trying hard to avoid batteries with my robot stuff because it is just a reoccurring expense and any "work" that I might get out of the robot would just end up with me having to pay the robot in batteries haha. But if I had a string of outdoor lights wired together in parallel and fiddled with its work week according to the available light I could probably just keep stringing together outdoor lights to keep what it can provide above what it requires to work.

That nonsense block was very helpful, babble on please! I'm pretty new to this and could use all the tips, tricks, and theory you can fit into the conversation, maybe later it will help me with something else.

With running a robot off a computer, my original vision was that I could just slap together some kind of breadboard computer have ram and fans and a harddrive connected to the robot itself , held together with some kind of playdoh and somehow it would work. As I went along I seen that doing that would require more programming than I could do so I would probably have to have the robot lug around a tower which to me sounds like a waste of energy (unless I converted the tower into some kind of addition function, like a platform for it to manipulate stuff on or whatever).

Where I'm at now though is, I think I want everything tethered and the brain part of it just sitting by the side, since everything would be tethered and I don't want stuff getting tangled and tripped on I'd probably need the wires to be suspended and to be careful with the programming to make sure tangles are avoided.

Since I'd be using a windows PC, would there be a way to get an embedded programming effect for robots (less tower but a containing what it needs to run) using program launching macros at the systems start up or would the compatibility just be too different?    I'm still hazy and maybe trying to cut too many corners.

Thanks for the offer to message you.   I might do that.

[mklrobo]

  Hello!
I would offer an opinion in regards to the power problem. 
If you use a laser, or ultraviolet power source, the solar cells
would have to charge on this frequency of light. Not too sure
how that would go. The flood light would be good, because
any camera processing done by the robots would need light
too.
I would suggest a battery exchange station. The robot has a battery
that gives so many ampere/hour work; then needs to recharge.
Have a battery exchange station, when the robot backs up to the station,
the station changes the battery, and also has ready charged batteries
waiting for the rest of the robots. It charges the batteries while the
other robots work. The station could also be a "control site" for communication
between the robots, to coordinate the work. The control site could relay
the progress of the work, any other information you find that you
will need, as well as house a main solar cell to aid in recharging.
Keep me posted..........