Robin, please do throw as much cold water as you feel is necessary. I'm not going to get offended if someone points out flaws in my ideas. On the contrary, I'm usually happy to be proven wrong, since that is how learning works.
You do bring up a good point about the devices potentially not liking being completely shut down and questioning what happens when they get the power back. However (assuming they don't mind being cut off and that they don't do some extra work once the power is reapplied), I believe that the probability of more than one device deciding to turn on at the same time (or close enough for the Arduino to not be able to recognize it in time and shut down the second one) is low enough to consider this approach. After all we are talking about fractions of seconds. I wasn't able to find a proper datasheet for that specific sensor I linked, but some others I was able to find have response times of 100 to 500 ms. I assumed the appliances described worked based on thermostat.
Where do you go from here? Well, if if no one else here suggest an alternative or smacks down my proposal because of some obvious flaw I am missing and you decide that the approach I described is suitable for you, your next step would be to determine if the devices can be "multiplexed" this way or if they need more operating time than available. If they can, then you would go get some basic supplies (an Arduino board, a breadboard, a few LEDs and resistors). Then you'd write the code and throw together a mockup of the project on a breadboard using some LEDs in place of relays and simulate the power usage with potentiometers or resistors or something. For the final device, you would certainly need some form of safe enclosure box to protect the user from the mains power. The box would have a power plug and as many sockets as you need (plus a few extra, just in case?). It would also contain the power supply for the Arduino, Arduino itself, the relays with a few extra components to drive them and finally the current sensors on each socket. It would probably be wise to start with a schematic, just to figure out how all the components will fit together and what exactly you'll need. You can also start looking at the relays available to you. Once you find them you'll have a better idea of what kind of driver for them you'll need.
Oh, I see. Here's the idea I have in my mind: You bring all the devices to the operating temp, one by one, then provide power to all the devices keeping them turned on. Since they are most of the time idling they are not drawing much current and your outlet should have no problems providing it. You then monitor each device's power consumption and as soon as one of them starts to draw more power than when idling, you cut the power completely to all the rest. You keep monitoring that device and as soon as it goes back to idle you start providing power to the rest of the devices, one by one, always checking if the newly powered device is still idling or needs to ramp up. If it idles you turn on the next. If it needs to work you cut the power to the rest. That way you achieve a sort of first come first serve system.
However, this can only function if the total working time of all devices is less than real time. If you have 5 devices, each requiring 10 minutes of power per hour (total 50 minutes), you can manage it, but if they each require 15 minutes per hour (total 75 minutes) you're out of luck.
It sounds possible, sure. You'd need a relay for each device you want to control, and some way of determining which device should be turned on. Since you mention operating temperature, I suppose a thermistor, thermocouple or something similar, depending on the temperature range you need to monitor and needed accuracy, will do the trick. Since you are switching mains at some significant currents and the relays will need to be substantial, I suppose some sort of driver (transistor, MOSFET) will be needed to drive the relays, as it is likely that the microcontroller will not be capable of suppling enough current to switch the relay directly.
No, you don't sound ridiculous, but every bit of information helps us in helping you. You mention upstream and downstream, so it's a river? Rivers I know of often have debris floating around. Your small boat could easily get entangled in a branch of something. If you navigate very close to the shore, there are even plants that live there. Also man made structures could easily protrude from the shore and disrupt your boat. Even if you anticipate that and add various sensors to avoid such obstacles, you will be hard pressed to filter them out of the final shore contour.
Are you sure the a boat is the best way to do what you want done - map the shore? Wouldn't an aerial photo be not only faster and more accurate, but cheaper as well? (hint: You don't need an airplane.)
When I first read the topic title I was going to direct you to the Microtransat Challenge, but after you mentioned "another sensor to ensure the depth stays above say a foot or a meter" I'm pretty sure that won't do. So, you'll have to answer a few questions and provide a bit more information. What kind of boat are we talking about (size, type...)? Power source? Autonomy (time and distance)?. Expected water surface conditions (shores of the Pacific are quite different from some small lake)? Why do you want it to follow the shore? Why do you want it to follow the shore so closely?
What I've noticed during shooting video (with Canon DSLRs, you too seem to be using) short flashes of light don't look nice in video. CMOS sensor in the camera doesn't actually take instantaneous pictures but rather scans gradually through the sensor, so one part of the picture always shows an earlier image than some other part. The effect is that with short flashes the frame tends to be partially exposed showing horizontal lines.
This is what I mean (around 1:45).
The flashing light in this case were too LED strips controlled by a DMX thingy.