The autonomy of modern smartphones is improving, but in general, battery technology has not evolved at the same pace as other components such as displays and processors.
Most of us need to charge our devices every night and, in some cases, spending a full day requires some planning or a mid-afternoon recharge.
Google improves the prediction of the time you have left battery analyzing your habits of use
Google has recently introduced smart battery statistics that estimate the remaining life of the battery based on smartphone usage patterns.
The feature seems to have been implemented last month, when Google uploaded Device Health Services, a system application, to the Play Store. Yesterday, a Google technician was seen by the Pixel product forums to explain how reasonable prediction of battery life works.
Previously, Google used simple assumptions to estimate battery life. If you used 10% of your battery in the first hour of use, you assumed that you would use 10% each hour. This process had little of intelligence, and for that reason, the battery estimates were always unrealistic.
Google has now adjusted that process with a bit of machine learning. The predictive model takes account of past use and uses it to predict future use. It takes into account the usage patterns in similar days and hours to determine the amount of battery you will use the rest of the day.
If you always watch Netflix or listen to a podcast when you return home. Also, take it into account and adjust the autonomy estimate accordingly.
In addition to giving you an estimate of the battery time, you have left. Also, Google shows you how the battery will decrease over time depending on your general use.
If you open the battery graph, you will see the expected battery life for the rest of the day. Now you will not see a straight line with the predicted descent, as in the image that heads the article. But you can see when Google foretells that you will use your phone more as the line descends more abruptly.
This feature seems to be active in the Pixel 2 and Nexus 6P, at least in those running Android 8.1 DP.