Hobby Tx are on continuously and then off-keyed for short ~50uS of pulses, according to old ARRL Handbook chapter about PPM. The frame separator or blanking time is longer circa 50 ms and since any interference would add to the strength there would be no glitches, even though the 50microsecond off pulses could be interfered causing missing pulses, for which the PPM modulator should be designed to tolerate, i.e. designed to reset if idle interval longer than the "blanking time" comes even if pulses are missing. That way a good set of PPM data next cycle can be received.
FM systems do not have the carrier power drop out, the frequency just shifts a bit to signal a mark. SNR is better since there is still carrier power and a interference source is still many dB below the carrier at least is the near range of the transmitter.
The PPM rxeris still robust because of the PPM will be repeated and the blanking time reset will bring the PPM back in sync in the event of an upset.
FM transmitters use xtal frequency synths and the vco is pulled to modulate. Rxer are narrow band FM receivers, single conversion from RF, IF limiters, and quad demodulators to get baseband.
So if the transmitter is sending continuous control data, then burst mode seems overcomplicated. WiFi and other packet radios deal with the issues of channel training, sync prambles because of the scattered bursty data and channel sharing issues.