What is the reason for the inaccurate monitoring during the sleep monitoring process?

Organize as follows:
The principle of sleep monitoring is as follows:
The sleep algorithm determines whether to fall asleep by detecting changes in the accelerometer sensor (ACC). When the ACC data changes little for a period of time, it can also be considered as a small amount of activity, and the algorithm will judge it as falling asleep; when the ACC data changes for a period of time, it can also be considered as a large amount of activity, and the algorithm will judge it as falling asleep . In scenes such as playing with mobile phones and watching movies, since the change of ACC in this scene will be relatively small, there will be a phenomenon of accidentally falling into sleep. If the sleep high-precision monitoring switch and sleep breathing quality monitoring switch are turned on, the data of the PPG sensor will be used, and the algorithm will analyze the sleep state through the ACC and PPG data, and obtain the sleep staging result and sleep score.
The reasons why sleep monitoring is inaccurate are as follows:
If the wearable device itself does not record sleep or the detection time is shorter than the actual sleep time, it may be that your wrist moves more during sleep, and the sensor determines that it is not sleep, resulting in no record or short sleep time.
If the sleep recorded by the wearable device itself is more than the actual sleep duration, it is possible that you have less wrist activity and are in a very quiet state when you are not sleeping (playing with a mobile phone before going to bed or sitting quietly watching a movie is probabilistically judged as falling asleep), resulting in The sensor is judged to be in sleep state.
Please check whether you are wearing a double-ring leather black and white strip. When using this strip, it is easy to cause the back shell of the strip to not directly contact the skin, which will affect the wearing detection and cause false detection during sleep and sleep.
Notice:
1: The algorithm side continues to optimize possible sleep missed and false detection scenarios, but limited by the algorithm detection principle, similar problems will exist in the industry.
2: For quiet scenes when not sleeping, promote the joint detection of multiple sensors and auxiliary equipment to further improve the detection accuracy.