The first step in this project is to build a test rig, a good start is choosing the hardware for the transmitter/receiver and MCU is to look at existing projects. I bought a couple of popular NRF24L01 wireless modules to test the range, later I'd use the SMD variant if the tests prove satisfactory for mounting on the PCB.

As far as the MCU is concerned I'm most confortable with AVR chips and I already have developing tools for it (USBasp programmer) so I'm leaning towards that. I know that the AVR MCUs are quite stable within a large input voltage range which is perfect for this kind of battery powered project, for example some of the ATtiny MCU can be powered with only 0.7V. They also have low power consumption mode theoretically allowing them to run for absurd amounts of time (years) on standard batteries.

Ideally no regulator has to be used, the NRF24L01 seems to run happily from 1.9 to 3.6V, but once the transmitter starts sending data it draws a lot of current from what I've read, an important voltage drop might cause problems when a battery with a high internal resistance is used (such as a coin battery), to be tested...

I'm now waiting for the delivery of the electronic components to start building the test rig. The first test is to check the transmitter/receiver range and if it can communicate between my mailbox/house.