Much clearer. Still, this is very ambitious for a digital person.
Of the 15 pins, only 5 of them do anything useful for your transmitter. So, you only have to transmit 5 signal. Those are your 3 colors, and horizontal and vertical sync. The color is impedance controlled and in the MHz bandwidth range. That will preclude you from transmitting RF without a license unless you are ISM. The TDMA switching you propose may not be the best or the simplest. You could consider matrixing the 3 colors into I/Q phase and amplitude. This may better control your phase shifting (time shifting) of the colors. But this is standard practice and nothing new. You do not have to treat all 5 signal equally in your TDMA scheme. You might want to send the sync's over digitally.
You mentioned sending the clock over a side channel. This will not solve your sync problem. At best you will have a phase locked clock, which is good, but not good enough. There are several good books on communication sync and slicers, many from Europe.
The analog mux you are asking about is quite simple and perhaps the easiest part of your system. Hittite (spelling) is one low switch vendor with the BW you want. However, this mux will drive your A/D cost very high. It is much simpler to A/D each line and digitally mux them. Eight bits each at 5 MHz is 40 Mbit per second for the color.
This might help on the 15 pins:
There are many good books on the USB protocol. You might try Jan Axelson first.
USB is bidirectional and is PC centric. The PC will poll your keyboard and mouse. If you use PS/2 for both you will eliminate the large overhead cost of USB.
Still PS/2 is a bi-directional serial protocol.
You not only have to send the signal, but send or know who is doing the talking at each moment. This will be tricky in your scheme.
Power needed:
< 300 mA keyboard, mouse even less. Can be powered through your USB connection.
Is this a hobby or a learning exercise? It is much cheaper to purchase these large system solutions. I would recommend learning how to do each technology on a small project. Then when you master each one, you can approach a large project. You could read the back issues of Circuit Cellar Inc. to get good ideas on each of these topics.
Best Regards,
John Solar