"We can see that in Ukraine, the air war has come extremely near to the ground, with low-flying missiles and drones targeting cities. A number of tools have been used to identify missiles and drones as early as possible. One of them is such a mobile-based object fixation - based on a way of thinking whereby practically every person could be seen as a detector," Col. Riivo Valge, the deputy commander of the Air Force, told ERR.

"Such a network will ensure that we are alerted as early as possible if a drone or missile is sneaking into our territory. A missile may be coming at 1,000 kilometers per hour, but drones may be coming at 100 to 150 kilometers per hour. This creates a buffer for us to take a notice and react," he said. Knowing about an attacking object earlier gives the air defense more time to prepare to shoot it down.

"Essentially, it's like a security blanket in which every man or woman is a sensor and the overview of items in the air is sent to the desks of those with defensive capabilities such as missiles, machine guns, or even airplanes to fire down the object," he said.

Airborne radars may not be able to see very low flying or small objects, so other means must be used to detect them.

Information as it comes in helps train the AI

The app's features are currently in a test phase, where photos and other data sent by people help train an artificial intelligence that analyzes the incoming information and helps airmen make decisions.

"The part we have in the phone is only the facade. The phone uploads the information to a server, where artificial intelligence sorts it into both reasonable and less plausible categories. Humans never study each report individually; artificial intelligence does it," Valge said.

The current request is for individuals to send in images of any flying objects via the app so that artificial intelligence could learn and train on additional data, according to the colonel. So it is not prohibited to send images of birds or other unrelated objects, for example.

"And it is not a problem if someone accidentally photographs a bird or something else. Likewise, it is not a person's concern whether the aircraft was friendly or hostile. You can photograph anything," he explained.

"Our engineers need baseline data. So we want to stress the machine: see how much it can digest, what the time reference is, and then optimize. Optimization, algorithm testing, and artificial intelligence needs an extensive amount of data," Valge said.

He said that the quality of the photo is not very important, it is enough to point the camera at the object. If you can't see the flying object in the dark or in fog, but you can hear it, you can take a photo in the direction of the sound: "To the best of your knowledge, you click away in the direction of the sound. This is also important information for us."

However, the software has been constructed such that if someone intentionally tampers with it, the images they send will be omitted from analysis, according to Valge. This is not currently used in the prototype, but the software has the ability to identify a person when used.

"Now, when we move out of the prototype phase – this will be truly one of the weapons of the EDF – there will also be a requirement for identification," he said.

The app was created by Estonian engineers

Valge also emphasized that although the idea to create such an app came from Ukraine, the app tested in Estonia is the work of Estonian engineers.

"Yes, the idea came from Ukraine, but it is entirely the work of Estonian engineers. Good Estonian men and women have tuned it and will continue to do so," he said.

The Defense Forces and the Defense League, notably the Air Force and the Women's Defense League, collaborated to develop the app.

This is not voluntary work, the creation of the app was paid by the state, the colonel said.

Prototype will run throughout May

The Air Force is inviting people to use the app starting May 10 during the Spring Storm exercise.

In the "Be prepared!" application, a sub-menu "Airborne object notification" opens, which can be clicked to take a picture of the airborne object. Once the photo has been taken, you can specify what type of object it is (or select: I'm not sure), then set the distance (50, 100 or 250 meters) and add a description if you wish.

The location and directional vector of the report, along with the image, are transmitted to the appropriate server, where all the data is combined to create an aerial image that increases situational awareness and improves civil defense, the Defense Forces explained.

The air threat notification function is "Be prepared!" will be available in the app from 10 May until 31 May, but the main test period is scheduled for 14-16 May, when the Spring Storm (Kevadtorm) exercise will be more active.

Objects spotted in the air can be reported from all over Estonia, but most of the reports are expected from the areas of the Spring Storm exercise, i.e. Pärnu County, Läänemaa, Viljandi County and Saaremaa. All aircraft sighted in the exercise area could be reported.

When the "Be prepared!" app is downloaded on the user's mobile phone and automatic updates are enabled, the app will automatically update with the Air Alert feature on May 10. If updates are not enabled, users will need to manually update the app or download it again from the App Store on Google Play.

--

Follow ERR News on Facebook and Twitter and never miss an update!