Hello, I'm thinking about smart home design, how whole system should work.
I noticed there are 'one module/main computer', for whole house rooms. But if that main module fails - functionality is disabled for every room.
Due this, I'm thinking to have such design: each room has it's own node. All nodes connects to one, main computer, and that computer receives info/sends commands to specified nodes.
By the way, all sensors would be wired-up.
So, node would be like single Arduino Zero or use STM32F1... for more ADC's channels. You could connect like 16 sensors(STM32F103 has 16 ADC channels) to each node: light, fire/smoke detector, PIR, .., and so on.
And the main computer... Probably Arduino Due won't have enough juice for it(big LCD, like 7" or even 10" screen), so it could be like Raspberry PI. Also, it could have GSM module, and of course it would be connected to internet.
Communication between nodes and main computer would be using RS485. It would be a 4-pin wire from each node to main computer(RS485 TX/RX, VCC(12V), GND).
In case main computer fails, each node can function without it(control light by light sensor, turn on something at specified time and so on). All you have to do is provide non-interrupted power supply to each node.
For power supply, it could be: one primary, and another secondary. If primary fails, secondary turns on. If electricity is gone, there is another battery backup, to keep whole system alive for couple of days. (It wont control your lights of course, thats just for each nodes and their sensors to be in standby mode. Although it would display temperature in every room and other things, that aren't controlled by power mains.)
So, what do you think? Any suggestions to improve such design? Thanks.