Musings on Chicken Coop Automation

​Should I integrate an Arduino or similar microcomputer into my chicken coup automation or just create a completely mechanical system?
If Microcomputing sounds like a good idea which system would work the best? Maybe a combination of Raspberry Pi and Arduino?
Lately, I have been contemplating methods of automating the process of keeping my chickens. I have a couple of specific plans for how to automate the food and water process, both of which are pretty simple, but there are other systems which require a technologically advanced approach. If you like building stuff or enjoy backyard automation, read on!
Watering System
For watering the chickens, I plan to use a simple 4”PVC Pipe which runs through the wall of my garage (which is a heated space) directly into the adjacent lean-to/chicken coup. Inside the coup, I will simply Put a t and run 2 feet of pipe out each direction. My hope is that even in the winter the water in the coup will stay thawed with the heat of the lamps, the body heat of the chickens, and the heat of the garage wall which is kept year round at about 45 degrees. The waterers that I plan to purchase are these:
These nipple waterers require less than 1 PSI of water pressure, so supplying a regulated constant water pressure of 1 PSI was my initial plan, but I have decided against this option, due to the cost of the pressure regulator. Instead, I will go with a solenoid water valve which can easily be electrically activated for filling the 4”pvc pipe. Each foot of PVC pipe has roughly .65 gal of water, so if I have 8 feet of pipe, that would be a volume of just over 5 gallons, which will last 5 birds probably at least a couple of days even in the heat of summer. The part where I am having a hard time is with some type of a switching mechanism. There are two options that I have considered.
1st Option- a small weighted float connected to a micro switch that will work somewhat like a sump pump valve.
2nd Option- simply run three wires into the water. One at a high point, two down below.  When the wires are exposed they will have no resistance between the wires, this input should be able to be calculated simply using the Arduino computer.
Food System
Food system, I don’t necessarily want to utilize a special loading system, but I do think that 4” Pvc pipe with an elbow at the end will work good as a food filling system. With that said, I do want these feeders to be capped off most of the time, only opening in the evening for the feed time, around sundown. I also would like to incorporate a photo voltaic sensor into each of the feeders, so presumably, they would allow the computer system to notify me when the food is low, because that would allow light in to the sensor.
Door System
From the research I have done, I think that the guillotine style door and rollers will likely be the best apparatus for opening and closing the door. Regardless of whether or not I integrate a microcomputer, I think that a simple photovoltaic switch and automotive window regulator attached to a chord should do the trick for this system. But it would be nice to be able to control these systems and check the status on via Wi-Fi on my smartphone, but, alas one project at a time.
The hardest of all- The chicken location tracker
When chickens get outside in the run, sometimes they can get carried away and not realize that it is time to come in to the coop for their nighttime rest and to get their evening snack. I would hate to have the door on the coop close a chicken or two out of the coop where they are safe from the many coyotes we have roaming our property at night. The only way that I can think to keep track of these birds is via RFID leg bands which are actually readily available. Initially, I thought that the combination of a simple RFID scanner in the doorway would allow you to know that all the chickens are inside. But, just because you have a scan of a bird just before the door were to close doesn’t guarantee that the bird is actually inside. The only method I have come up with which might work is to make their entrance/exit a bit of a tunnel, where the chickens would have to pass through two sensors with which the computer could possibly sort out the location of the bird based on the order of the scan.
 
I have many other ideas as well… A predator sensor, activating a light and letting me know in the house. Possibly an RFID system to identify birds that are spending most time inside or outside, which could help me to identify birds likely to go broody. Finally, temperature, lights, and moisture control within the same system would be awesome.
If any of you have more ideas, comment below!