Earthquake Early Warning based on smartphones Earthquake Early Warning (EEW) systems involve very high Capex. Therefore, EEW systems have been installed in a very limited number of regions. Applying crowdsourcing on humanitarian projects led us to use accelerometers in smartphones distributed all over the globe to detect earthquake shakes. We have implemented Earling, an innovative method of EEW system, utilizing personal smartphone sensors, to detect earthquakes by gathering shake reports. To minimize false alarms, we have devised some special algorithms, called Trust Algorithms (TA). Data sent from smartphones, is analyzed by TA in real time. In communicating with a service, the system can receive shake data from smartphones, analyze and distinguish earthquakes from everyday shakes. Then central unit applies complementary analysis on the results and send an alert to users in the vicinity of the earthquake in seconds. We have successfully tested some of the recent earthquakes using simulation data around Europe by Earling, at first to detect the earthquakes and then to issue an alert in simulated environment. Using algorithms embedded in Earling, we can detect epicenter and magnitude of earthquakes, as well as microseisms. Therefore, detecting a massive earthquake in the coastal region means a tsunami may be on its way. Earling can act as a complementary system for traditional early warning networks and following local public-safety policies, issue a tsunami early warning alert, to areas at risk.
Earthquake Early Warning System based on smartphones Project Earling 7 th Tsunami Symposium JRC, Ispra, Italy Farzad Azima Circular World Foundation fb.com/acircularworld
Earthquakes August-September 2016 3
Earthquake Impact 400,000 Earthquake Impact 2000-2016 350,000 300,000 250,000 200,000 150,000 100,000 50,000 0 2001 2002 2003 2004 2005 2006 2008 2009 2010 2011 2012 2013 2014 2015 2016 Casualties 4 Injured
One Second is Matter 5
Regions under cover of EEW Countries including Japan, Taiwan, Mexico, Turkey and Romania currently have operating EEW systems. These systems are generally government operated and funded. (Moore, 2016) Some other systems also being used to alert test users in California, Oregon and Washington, Chile, Israel and Korea 6
Speed of Earthquake waves http://earthquake.usgs.gov/learn/topics/seismology/keeping_track.php 7
Project context and challenges Utilizing smartphones to detect earthquakes Huge amount of data was expected to being sent to servers Avoid false positive and false negative alerts Distribution of the client side which is an app in different countries and different level of frightening against earthquakes Trust 8
Real Earthquake in California Simulated -116.3 34.1 34.2 34.3 34.4 34.5 34.6 34.7 34.8 34.9-116.35-116.4-116.45-116.5-116.55-116.6
Real Earthquake in Italy 15.3 Simulated 15.25 15.2 15.15 15.1 15.05 15 14.95 14.9 14.85 14.8 41.85 41.9 41.95 42 42.05 42.1 42.15 42.2 42.25
Earling Client Side 11
Current Distribution of Earling
Algorithms of Earling Detect smartphone shakes by embedded gyroscopes Distinguish earthquakes from everyday shakes Estimate direction of earthquake waves Estimate epicenter of detected earthquake Estimate vicinity areas and issue alert in few seconds
Coverage of Chile 14
Coverage of Puerto Rico 15
Coverage of Ecuador 16
Coverage of South America 17
Earling in High Risk Areas Active nodes on 10 September 2016 18
Architecture of Earling
Earling Client Side Diagram
Earling Server Side Diagram
Earling - Client Side Operations Receive shakes Apply elementary & lightweight algorithms Keep connected with servers Send report Receive alert 22
Earling - Server Side Operations Provide network infrastructure Analyze retrieved events from smartphones User authentication Trust Algorithms & Fake Signal Detection Apply decision algorithms to issue alert Resolve BTS antennas to geo-location Issue alert 23
Earling Client/ Server Client Side Gyroscope Earthquake Detection Connection Keeper Server Side Analyze Events User Authorization Fake Signal Detection Decision Algorithms Alert Receiver Reporter Connection Keeper Alert Receiver Reporter Resolve User Location Issue Alert 24
Technical Challenges 25
Recorded Earthquakes by Earling - Chile M5.1 21 July 2016 26
Recorded Earthquakes by Earling - Chile M6.1 25 July 2016 27
Recorded Earthquakes by Earling - Chile M4.3 28 July 2016 28
Recorded Earthquakes by Earling - Ecuador M4.7 26 Aug 2016 29
Crowdsourcing Manage hundreds of thousand of human resources Earling designed a game Send report Receive alert Give something or do something
Feedback Analyze user comments Analyze sent alerts Following local public policies Algorithms improvement Improve reliability Development 31
Alert No early warning has been issued yet Due to ongoing improvement on Trust Algorithms and measurement of the areas should be noticed Lack of the number of required sensors in the crisis area Not facing with huge shakes the coverage area
Limitation Timing False alarms Estimate epicenter Estimate magnitude Victim area 33
Future of Earling Tsunami Early Warning Mesh of WiFi Gathering automated statistical data Integration with other systems API for developers Network operator based alert Training (Creating mind map) ios, Windows and Blackberry 34
Distribution Steps Phase 1 Country Priority Country Priority Chile 1 Spain 5 Ecuador 1 Italy 6 Costa Rica 2 Philippines 7 El Salvador 2 Indonesia 8 Mexico 3 Turkey 9 Nicaragua 3 Pakistan 10 Greece 4 New Guinea 11
Distribution Steps Phase 2 Country Priority Country Priority Oklahoma, US - Alaska - Taiwan - Tajikistan - New Zealand - China - Fiji - Iceland - Peru - Japan - Panama - Nepal - California, US - Russia -
Facilities Can Speedup Scientific support Following guidelines of millions of installations Advertising, Social Media Governmental and UN level support Private firm cooperation
Circular World fb.com/acircularworld farzad.azima@outlook.com