Had to break this up - I'm writing a novel here!
;D
Regarding the steering motor (and needed controller): It shouldn't consume as much amperage as the drive motor, but it may, so keep that in mind so you can build or buy the controller needed.
Something else: Why am I telling you to buy motor controllers? Because you don't have the skills yet to build such controllers (neither do I!) - and you likely waste more money trying to build one (and learning the hard way) versus just spending the money on a controller. Fortunately, a controller for the steering motor or linear actuator likely won't need to be as beefy as the one for the drive motor, so there's that. But you need to be aware of the specs of the motor (stall/running current needs), and design accordingly.
- Once I have steering I want to look into unmanned or at least remote control. Being familiar with a Roomba i.e. (The "it doesn't know where it is, it only cares if it bumps into something and the sensor tells it to back up and turn slightly and try again" part.) I know it does more than that to cover areas decently but for now I just want to emulate that part.
You said you were going to tow the mower behind the 3-wheel scooter, right? Have you thought about how you would get such a thing to back up properly? Unless I am misunderstanding you, you are going after a trailer-style linkage, right? So to back up, you would have to somehow get the robot platform to understand how to "back up a trailer", just like a driver would...
If you can figure out how to automate that, you will have an invention worthy of the name!
3a. For unmanned control I was thinking of four posts to quadrant off an area with some type of beam (light, laser, sound, etc.) then when the mower hits the beam it stops and reverse x amount of time/distance then turns and tries again until it has free movement. I will start out with squares for now so that this is simpler. In effect using the beams as the bumpers/sensors.
This probably won't be reliable to detect or such, or you will have false positives, or you will need some kind of interface from the beam-break (RF or such) to the robot. I understand what you are trying to get at (ie, getting the robot to "see" or "hear" the beam of light or sound it is "breaking"), but such a scheme likely won't be reliable...
I would instead suggest a buried or staked-down perimeter fence wire, which you run a modulated and coded signal through using an amplifier, the output of which can be picked up by the robot. One of these:
...suitably hacked could certain work, and its cheap enough.
Alternatively, you could bury magnets spaced so far apart along the perimeter, then use a hall-effect sensor to detect them...
3b. For remote control I was thinking of video fed to my iPhone/iPad and then I can steer it myself from in the house. I actually suffer from heat intolerance severely and can die if exposed too long to the sun in hot weather so this would be great to cut the grass from an air conditioned environment. So far I know that the remote control can be done with some combination of OSC and maybe xBee or something for controlling and I thought I saw a video board but haven't researched that yet.
I would actually use something like 802.11 wifi - something like this Arduino:
http://www.robotshop.com/productinfo.aspx?pc=RB-Asy-02
So, I really would appreciate any suggestions on number 2 and 3a. I'm going to document each step of this project. Even though I think this is a lot to shew on being a total noob I am not duanted by the task and plan on doing the best I can to reach all of these goals.
Any help and suggestions are greatly appreciated.
Though it may not seem like it, I have barely scratched the surface of what this project is going to entail for you to build, at least properly for a safe operating device. Given its size and what you plan on doing with it, plus the need/want for remote control, I once again want to impress upon you the importance of a safety system. The last thing you want is a small child or dog running into this machine's path, and having no way to stop it.
With that said, there's a lot to do here, if you want to do it right without throwing good money after bad. It won't be cheap, but at least you have the most expensive part out of the way (the scooter - I'm sure you know how expensive they can be, even used). Take your time on the project, design, plan, and build one component at a time, and it should work out. I would also suggest bench-testing using smaller components and/or "stand in" systems; if you can build a model of the system (maybe a scaled down version), it might help to work out some of the control system bugs without danger of injury or other destruction while testing.
Good luck, and I hope it turns out successfully for you!
