Hi,
My question is regarding the FCC requirement that each frequency must be used equally on the average by each transmitter.
Consider a synchronized system where 50 channels are used. A predefined sequence of 50 randomly ordered channels are employed. One slot is 10 ms @ one frequency. So the 50 channel sequence is repeated every 500 ms.
A node in the system is triggered at a random point in time and would like to transmit its packet as soon as possible, i.e. there is a maximum of 10 ms delay until the start of the slot.
One of the rules states that the maximum dwell time is 0.4 sec over a 20 sec period - this requirement can easily be withheld by using an moving bucket algorithm ensuring that each channel will be transmitting for no more than 0.4 sec accumulated time within the last 20 seconds.
However, the requirement that each channel must be used equally on the average is a bit more difficult to interpret with respect to implementation.
Let's say that the node is triggered, each time, when the channel sequence is at channel 35 (this is highly unlikely though due to the random nature of the trigger event, but is still possible). Is the FCC requirement still obeyed - since the nature of the trigger event still is random? And that the channel still is picked from a list of 50 frequencies?
One other solution could be to define a flag/counter for each channel, and only use channels where the flag has not been set yet. When all 50 flags have been set, all flags can be cleared and the algorithm starts over again. This would lead to long waiting periods for the triggers and the packet would not be sent as soon as possible. A better way perhaps could be to use the counter approach and set some requirement here.
So my question is really what is meant by that each frequency must be used equally on the average - explained in a more precise way for a synchronized network with random events in time which is to be transmitted.
HC