Was that a tremor?

IoT powers earthquake early-warning system

There’s a 62 percent chance of an earthquake in the San Francisco Bay Area in the next 30 years, according to the U.S. Geological Service. But no one knows when or where the next big one might be. So, when I heard of startup offering an early-warning service for earthquakes, I wanted to learn more.

It’s a tough problem to solve. Thousands of widely dispersed sensors are needed, as an earthquake could originate anywhere. To be effective, an early-warning service also needs a dependable communication network and fast analysis. Take a look at what happened when the last "big one" struck the San Francisco Bay Area.

If there were a person meant to develop such a solution, it has to be Battlagazi Yildirim. Born in Turkey, he personally witnessed the devastation caused by earthquakes. This motivated him to study math, engineering and geophysics. He earned his PhD from Brown University and did post-doctoral geophysics research at Stanford University. With a $1 million grant from Verizon, he founded Zizmos to provide an earthquake early-warning system that uses sensors, mobile apps and cloud-based services.

Tremors radiate outwards from the epicenter of an earthquake at the rate of about 2 miles/second. So, a tremor from an earthquake centered near Santa Cruz, California, would take about 30 seconds to travel 60 miles north to San Francisco. By collecting and processing tremor data within five seconds, residents in the earthquake's path can be given about 25 seconds advance notice. This is usually enough to evacuate a house, close a bridge or shut down computer servers. Earlier approaches that relied on dedicated sensors were often too expensive or didn't cover enough area to be helpful. Battalgazi decided to take a different approach, leveraging the sophisticated sensors already in smart phones.

The sensors
Zizmos uses the accelerometers in smartphones to sense tremors and integrates them with its free iOS and Android mobile apps. The mobile apps serve a dual purpose in that they also serve as the alert in the event of a tremor. Thousands of sensors are needed in dispersed locations for the service to be effective, as the epicenter could be anywhere. The Zizmos mobile app is optimized to sense tremors when the phone has Wi-Fi connectivity, is stationary and connected to a power source. The app then sends the vibration readings to the cloud for analysis.

Data analysis
False-positive vibration readings (such as from a nail being hammered into a wall) are filtered from the collected data. Zizmos then applies data analytics to triangulate the epicenter location and direction of tremors to determine who might be impacted. Sensor data is collected, filtered and analyzed quickly to maximize the advance notice provided. Everything is optimized for speed, as every millisecond counts.



Alerting
Since an alert may have to be sent at any time, Zizmos decided to use a "push" notification mechanism rather than relying on users to have their Zizmos app running. This enables the app to maintain a direct and persistent cellular or Wi-Fi connection through which an alert can be sent. They use Apple Push Notification service (APNs) on iOS and Google Cloud Messaging (GCM) Services on Android.

Do your part! Install the free Zizmos app before the next "big one."

--
Deepak Puri is the Chief Solutions Architect at SkilledAnalysts.com, a San Francisco-based IoT consulting firm that helps clients develop Internet of Things (IoT) prototypes. Clients reduce their risks and expedite projects by evaluating an IoT proof of concept (POC) before making any major investment.