Today, hundreds of civil drone models are available for sale — from fairly simple toys to impressive aircraft built to carry substantial weight. Most of them are used for entertainment or aerial video. But, as is the case with almost all technologies, sooner or later someone puts them to malicious use. That creates demand for systems to get rid of unwelcome aerial presence. Demand, in turn, breeds supply, and our enthusiasts have developed a solution called Kaspersky Antidrone.
How can one be wronged by a drone?
To begin with, any civil UAV is, above all, a remotely controlled camera. Naturally, the world has already seen reports of drones used by burglars to identify potential targets and research defense systems; by headline hunters to get footage of film sets for then-unreleased episodes of Game of Thrones; and by some strange individuals to peep into the private lives of St. Louis residents.
Drones can also carry loads and, therefore, serve as smuggling vehicles. True, their weight-lifting ability remains wanting so far, but you don’t need much of that as far as certain substances go. In fact, according to The Washington Times, Mexican drug gangs use drones extensively to smuggle drugs across the U.S. border. Potentially, drones can also be used to steal expensive materials (such as platinum or diamonds) from guarded facilities.
And finally, a drone is a piece of metal that may suddenly crash into something. I am not even referring to intentional ramming attacks. There are many reports of drone accidents that have led to rather unwelcome consequences.
How to combat drones
Of course, we were not the first to come up with an idea for a drone repellent. Other companies have proposed a variety of UAV response methods. Some of them offer detection systems that only warn the user about approaching drones, but others try to combat this aerial threat more actively.
Some have suggested the use of nets; others would train predatory birds to serve as interceptors. Some systems use a fake GPS signal to disorient the device. But most antidrone solutions represent variations on an electromagnetic cannon designed to bring the drone down, which we believe is fundamentally wrong — not to mention, not always legal — just like attempts to hack communications between the apparatus and its remote control unit. We prefer the most noninvasive method of all: targeted suppression of communications between the remote control unit and the drone.
As soon as the control signal is lost, a normal UAV will immediately initiate the landing procedure or try to return to the launch site. Some models are programmed to hover or fly to a different location in such instances, but these are rather an exception to the rule.
But landing a drone is only half the battle. At the heart of our solution lies a system for automatic detection and classification of aerial objects.
The principle behind our method
As you might guess, we do not manufacture any jamming or detection hardware. We are a software company. Therefore, our experts have written software that uses data from various sources to detect drone approach, aim jamming antennas at it, and force it to land. The solution is adapted to the customer’s actual needs and configuration by our system integration partners.
To understand that a UAV is approaching the protected area, the primary detection module is used. A detection module is virtually any device able to scan the airspace over your perimeter — a camera, lidar, radar, or an acoustic sensor. In fact, even IPTV cameras can feed data to our system (provided they cover enough sky).
Each method has pros and cons. Our demo setup uses a lidar — a device that detects objects and measures distance using monochromatic light emission (in this case, a laser). True, it can detect an airborne object at a shorter distance than, say, radar. But it can also tell a drone from a bird from much farther away. It also is not subject to specialized licensing (in some countries it is next to impossible to obtain approval for the use of radar in a populated area; radar emits radio signals that may conflict with other equipment).
Once it has seen (or heard) aerial presence, the primary detection module transmits the object’s approximate coordinates to the server and the classification-and-jamming module. This module directs a high-accuracy camera toward the flying object and gets it in focus. The image is sent to the server, where the actual classification takes place.
Thanks to machine-learning algorithms, our system can not only distinguish between birds and drones, but even determine the actual drone model. As soon as the program confirms the object to be a UAV, it activates targeted noise generation. True enough, attackers may attempt to disguise their drone, but then it will simply be classified as a UFO — and the system will respond to a UFO the way it is programmed by the user. To be on the safe side, it is better to jam across the entire available frequency range.
Basically, everything depends on the needs and resources of the customer, as well as the integrator. The essential components are the jamming equipment covering the frequencies used by drone manufacturers, a server to run our software, and a high-accuracy camera. Any primary detection devices can be used, and they can be combined to improve detection accuracy or range.
One jamming module can track one target. This is normally more than enough, but in case you anticipate being attacked by flocks of drones, the system is quite easily scalable. The server simply needs to be powerful enough to process the incoming data.
If needed, a mobile solution can be designed to be mounted on a car — not to operate while on the move (this is possible, but it would require a minor software update, and so far we have not seen any demand for this option), but for field use. This is exactly the modification to respond to the movie industry’s needs to prevent unauthorized film set photos.