Android_Battery_Prediction

This Android application is designed to run a set of workload tests, and show the current battery levels of the device. It runs a set of workload tests from time to time and

generates a test report. The test report contains the details regarding the current battery

status.

The idea behind this application was to make battery prediction accurate. Instead of just

reporting the current battery and extrapolating the expected time until its runs out, this

app simulates a normal usage and then predicts the battery based on the simulated

usage.

Introduction

The workload tests are designed to simulate a general usage of the device by an average

user. The workload test includes the following actions:

1. Launching a random website in a new tab of an existing browser instance

2. Launching Google Maps and showing the current location on the map

3. Launching and running the execution of a random application on the phone using

the MonkeyRunner api for Android.

These workload tests are performed repeatedly in fixed time intervals. The battery

consumption is reported in the form of a graph which plots the Output Voltage of the

battery.

Tools Used

Eclipse v3.8

Android Development Tools v22.3.0

Emulator Nexus-4 AVD

AndoidPlot library 0.6.0

MonkeyRunner API

User Interface

The user interface displays all the following data:

• Current level in %

• Expected duration the battery will last

• Current battery temperature in C

• CPU free %

• Current battery output current in mA

• No. of processor cores in the device

• WiFi status

• Bluetooth status

• Time of last workload test

It also displays a graph of the battery voltage over the course of the workload tests.

Principle

The principle behind battery prediction is to divide the battery capacity in mAh by the

average current in mA over the course of the workload to produce the expected time. But

it isn’t possible to explicitly know the current delivered by the battery. After several

experiments and help from various other groups, I found out that there is no way of

getting Battery Current value through software only (as its not supported in h/w). Only

way I found was to measure current flowing through battery by means of multimeter.

This method is not useful for this purpose. Instead, I use the principle that internal

resistance of a battery is constant, specifically for Li-On batteries, which power all mobile

devices today[1]. This means the current changes proportional to the voltage. Therefore, I

use the voltage to predict battery life. The output of the battery during the course of the

workload is stored in an arraylist. The average value of the voltages stored in the

arraylist is used in the computation. The capacity of the battery is captured by obtaining

the manufacturer and model of the device, and then using the known battery capacities

of devices.

Program Details

The BatteryManager class

(http://developer.android.com/reference/android/os/BatteryManager.html) provides

the various functions to extract data regarding the battery.

The following extras of the BatterManager class are used

• EXTRA_VOLTAGE

• EXTRA_LEVEL

• EXTRA_SCALE

• EXTRA_TEMPERATURE

• EXTRA_VOLTAGE

• EXTRA_ICON_SMALL

Nested timers are used to run the workload tests. An outer timer triggers every 45

seconds. The inner countdown timer counts down from 10 and launches a new tab in the

browser when it reaches 0. The link opened in the browser alternates between a link

from a random set of links and a link to Google Maps.

AndroidPlot is used to plot the required graph (http://androidplot.com/). AndroidPlot is

an API for creating dynamic and static charts within your Android application. It was

designed from the ground up for the Android platform, is compatible with all versions of

Android from 1.6 onward and is used by over 500 apps on the market today.

The information from the EXTRA_VOLTAGE is stored in an arraylist. This arraylist is then

plotted on the graph using the AndroidPlot library.

MonkeyRunner API

The MonkeyRunner tool provides an API for writing programs that control an Android device

or emulator from outside of Android code. With MonkeyRunner, you can write a Python

program that installs an Android application or test package, runs it, sends keystrokes to it,

takes screenshots of its user interface, and stores screenshots on the workstation.

The MonkeyRunner script I have written performs the following actions:

1. Clicks the Home button

2. Clicks the Apps button

3. Clicks on a random app on the screen

4. Waits for 5 seconds

5. Takes a screenshot

6. Clicks on the Home button

7. Clicks on the Apps button

8. Clicks on the Battery Prediction app

The MoneyRunner script is executed on the terminal using the Python command. The script

is executed on the Emulator as well as the device (if connected).