RESUMEN
A mental and physical recovery after an awakening moment depends not only on the overall sleep duration and quality but mostly on the sleep stage in the waking moment. The most comfortable awakening moment is during the Light or Wake sleep stages. But the fix-time alarm clock doesn't take into account the sleep stage in the awakening moment, which often results in awakening during the Deep or Rapid Eyes Movement stages. To reduce the negative recovery effect, big companies and research groups develop various awakening systems. Such systems recognize sleep stages based on wearable sensors' data (mostly from accelerometer sensors) and thus can find the easiest awakening moment time with minimal recovery effects.However, it is quite hard to measure and verify the efficiency of such systems without using polysomnography (which can be performed only in clinical conditions by medical experts). To solve this problem we developed a methodology based on questionnaires and psychological tests. Such an approach has big scalability, does not require special medical equipment, and can be evaluated in a home environment with minimal support effort. The proposed verification approach has been tested on smartwatches with the sleep stages forecast model. The proposed model accuracy was 78%. Results of our experiment show that the majority of users demonstrated a correlation between awakening quality and the verification tests performance.