FREMONT, CA: Researchers at prominent universities have used machine learning (ML) to program an algorithm that can accurately predict a battery’s performance and eventually help accelerate its performance. Based on millions of measurements gathered from batteries—including charging times, power capacity, and even the temperatures of the cells—the artificial intelligence (AI) system can predict how many cycles it can last.
Frequent harvest of online data keeps smartphone users up-to-date, but it can also consume a large battery percentage. When people perform tasks on the internet, an automatic download may hamper network bandwidth, creating a negative user experience.
Smartphones possess a variety of data sources and sensors that supply circumstantial information to deepen our comprehension of patterns of phone usage through Machine Learning techniques. Moreover, on personal smartphones, the AI system should be energy-efficient.
Smart Battery Usage:
Power performance is a crucial issue for any system on smartphones or other mobile devices as users cannot charge them at all times. Instead of using power-consuming sensors, the AI system relies on software-generated data only. The predictions are based upon applications accessed and used, time of usage, and period of performance that has a negligible impact on the battery as shown by experimental measurements.
Positive vs. Negative:
Sometimes users unlock their phones to check on text messages, social media notifications, and phone calls, which are random and arbitrary, and difficult for AI to predict. But the same AI holds considerable promise regarding regular and definite events when it can minimize unnecessary energy consumption when people are not expected to use their phones.
During the non-usage periods, the AI in smartphones could schedule tasks like application updates, creating a better user-experience, conserving more battery life throughout the day.