An adaptive user-guided tool for power management of mobile devices - Embedded.com

An adaptive user-guided tool for power management of mobile devices

The battery life of mobile devices is one of their most important resources. However, due to battery size constraints, the amount of energy stored in these devices is limited. Many factors can impact battery life. Resource utilization by applications running on the platform and the number of powered- on device components the platform can greatly impact battery life.

As a result, the platform’s power management layer can change the processor frequency or suspend the hard disk in response to utilization. In addition, it can change the device components’ power states to an idle sleep state in an attempt to reduce the power consumption.

There is much research on power profiling of device components or energy profiling of applications in order to enable application developers to debug their applications from an energy-efficiency perspective. However, there is a lack of focus on the end user.

What if a user needs the platform to last for a specific duration until a particular task is performed, but the battery life is not enough? Can we guide users by giving them options to reach their goal? Will users be willing to completely sacrifice some options in order to achieve their goals?

By considering the mobile device as a provider of a collection of resources—similar to a cloud resource provider, which enables users to reconfigure the platform in order to include only the needed resources in order to achieve their goals and completely shut off everything else—then, yes, extending the overall battery life of a mobile device in order to complete a specific task will be possible.

As a result, we developed BatteryExtender, a user-guided tool for power management of mobile devices. It can predict the impact of applications and device components on a platform’s overall battery life through minimal energy profiling thus minimizing the power consumption overhead of the tool.

It consists of five components: calibration, user interactive (user command selection and user interface), energy profiling, power management, and monitoring modules. The calibration module aims to profile the power consumption of platform device components and save it to a configuration file. Using the user interactive module, users can enter the duration by which they want to extend battery life.

The selection triggers the energy profiling module that determines the list of applications running on the platform and calculates the estimate of their minimal energy consump- tion over the battery life and ranks the top 5 most energy- consuming applications. It also determines each device com- ponent power state (on or off) and active state (e.g, Bluetooth is actively connected to a device and transmitting or receiving data).

Then, it estimates the impact of changing the device component power state on the platform’s battery life. Upon completion, the energy profiling module updates the user interface with the top 5 power-consuming applications and displays the current active state of the device components in addition to displaying the amount of battery life saved/gained by changing their power state. At this point, the user can select the options to change.

Upon option selection confirmation, the power management module re-configures the device components in order to satisfy the user’s choices and termi- nate the check-marked applications. Upon completion, the monitoring module periodically checks the remaining battery life. The goal is to ensure that the remaining battery life satisfies the minimum between battery life extension duration requested by the user and the sum of estimated battery ex- tension duration based on the user’s selection.

Since the remaining life duration is not accurate, the monitoring module allows few unsatisfactory estimate readings. However, upon reaching a threshold, the energy profiling module is triggered again. Then, it is up to the user to either reconfigure the plat- form or accept the new expected battery life.

It predicts the battery life savings based on the new configura- tion, in addition to predicting the impact of running applications on the battery life. Through our experimental analysis, BatteryExtender decreased the energy consumption between 10.03% and 20.21%, and in rare cases by up to 72.83%. The accuracy rate ranged between 92.37% and 99.72%.

To read this external content in full, download the complete paper fron the open online archives at Wayne State University.

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.