sms fence detection

13
SMSFenceDetection App A Better Approach to Geofence Detection Mobile Systems - Master Degree in Computer Science, University of Bologna - A.Y. 2015/2016 Andrea Sghedoni - MATR. 0000736038

Upload: andrea-sghedoni

Post on 21-Jan-2018

92 views

Category:

Technology


1 download

TRANSCRIPT

Page 1: Sms fence detection

SMSFenceDetection AppA Better Approach to Geofence

Detection

Mobile Systems - Master Degree in Computer Science, University of Bologna - A.Y. 2015/2016Andrea Sghedoni - MATR. 0000736038

Page 2: Sms fence detection

Summary● Introduction● Environment● Challange● Strategies

○ Polling○ Auto-adaptive approach

● Testing phase● Performance evaluation● Conclusions● Bibliography

Page 3: Sms fence detection

Introduction● What is SMSFenceDetection?● Location-based Service App → Geofence

ENTER

EXITAlert SMS!

Mom’s phone

Page 4: Sms fence detection

Environment● Android SO● IDE Android Studio 1.5.1● GoogleMap Android APIs● ACCESS_FINE_LOCATION, ACCESS_COARSE_LOCATION and SEND_SMS in

AndroidManifest.xml● MinSdkVersion: 14 → Android 4.0 ICE CREAM SANDWICH● TargetSdkVersion: 23 → Android 6.0 MARSHMALLOW

Page 5: Sms fence detection

Challange● Trade-off between efficiency and battery consumption● LBSs may drain battery in few hours

○ need to implement a smart approach for background services

Page 6: Sms fence detection

Strategies● Background Android Services● Goals:

○ Geofence detection○ Send SMS Alert

● Strategies:○ Polling○ Auto-adaptive approach

Page 7: Sms fence detection

Polling● Every 5 sec● PRIORITY_HIGH_ACCURACY● Constant use of GPS Sensor

Page 8: Sms fence detection

Auto-adaptive approach● Expensive resources(as GPS Sensor) only when absolutely

necessary● Parameters:

○ Distance location-fence (0 - 5)○ Direction of device (0 - 5)○ Speed of device (0 - 5)

Page 9: Sms fence detection

Total Evaluation

Example 1:

Distance: 6km

Direction: yes

Speed: 110 km/h

fi = 5 * 0,8 + 5 * 0,1 + 4 * 0,1 = 4,9 → 5Result:

● Interval: 5 sec ● Accuracy: PRIORITY_HIGH_ACCURACY

Example 2:

Distance: 300km

Direction: no

Speed: 20 km/h

fi = 0 * 0,8 + 5 * 0 + 4 * 0,1 = 0,4 → 0Result:

● Interval: 480 sec (8 min)● Accuracy: PRIORITY_LOW_POWER

Page 10: Sms fence detection

Testing phase● Polling strategy vs Auto-adaptive approach

○ Period: 20h○ Distance:

■ 300 km, 10 km, 1 km■ movement(ENTER/EXIT events) in a radius of 2 km

● Measures:○ mAh consumption○ % battery consumption○ GPS utilization time

● Test device:○ LG Sprint with Android 6.0○ Li-Ion battery 2100 mAh

Page 11: Sms fence detection

Performance Evaluation

Page 12: Sms fence detection

Conclusions● An intelligent approach can increase energy efficiency,

with important results● A simple app, with a simple background location polling,

may drain the battery in a few hours● Auto-adaptive approach consumes a low energy level, so

that user can keep alive the background service for a long time

Page 13: Sms fence detection

Bibliography● Axel Kpper, Ulrich Bareth, Behrend Freese “Geofencing and

Background Tracking The Next Features in LBSs”, Berlin, 2011● Axel Kpper, Ulrich Bareth “Energy-Efficient Position Tracking in

Proactive Location-Based Services for Smartphone Environments”, Munich, 2011

● Creating and Monitoring Geofences, Google Android Documentation, 2016.

● Receiving Location Updates, Google Android Documentation, 2016.

● https://play.google.com/store/apps/details?id=com.sm.app.wifi.sghedoni.andrea.wifism , App on Google Play Store.