There a lot of ways how some indicators can be made adaptive instead of calculating fixed periods
One, less known method, is to use normalized ATR (Average True Range) for making the calculation adaptive. And since EMA (Exponential Moving Average) is a a good candidate for being adaptive (it allows fractional periods for calculation), here is EMA that is using ATR for adaptive EMA calculations
PS: as a comparison what does the adapting do to EMA, here is a comparison of the ATR adaptive EMA (colored line) and regular EMA (gray line), using same parameters. It is obvious that the adaptive is “faster” in the times of high volatility, and that way it is fulfilling one of the usual goals of making any indicator adaptive: to react faster when needed, and to become “slow” when the market is slow too.