Researchers from IIT Madras (Indian Institute of Technology Madras) have successfully designed an AI-powered drone that can take down rogue drones non-invasively. Developed primarily for law enforcement agencies and the armed forces, this drone can detect unregulated drones in the nearby airspace and proceed to take control of its flight. At present, there have been several changes in the Indian drone industry. In January 2020, nearly 20,000 drones have been registered with the DGCA. Furthermore, India has an estimated 6 lakh unregulated drones. Thus, an anti-drone system is the need of the hour.
The AI drone was designed by a team comprising Vasu Gupta, a final year BTech student, Department of Aerospace Engineering, IIT Madras, and Rishabh Vashistha, a Project Associate working in RAFT Lab (Department of Aerospace Engineering, IIT Madras). The researchers head a startup called Rodella that is also based out of IIT Madras.
Scope of the project
“The programmable nature of our aerial vehicles also opens up the possibility of swarming multiple vehicles to act as a team and accomplish a common mission,” stated Dr. Ranjith Mohan, Assistant Professor, Department of Aerospace Engineering, IIT Madras.
The Professor also stated that the current prototype can detect and visually track objects. Another key highlight of this drone is that it can be operated over the internet. Thereby, it can take down rogue drones without having to stay in the visual line of sight. Since the drone is controlled over the internet, it has another advantage apart from BVLOS; operators can send in a swarm of drones that can track and detect people, objects and other rogue drones.
“Our next step will be to conduct exhaustive tests on the system and ensure its reliability for catering to a wide range of demanding missions that pose a challenge to our law enforcement and defense agencies,” he added.
How does the drone work?
The anti-rogue drone is unlike other anti-drone systems. It tracks objects visually using Deep Neural Networks (Artificial Intelligence). The AI-powered motion detection algorithms enable the drone to track objects even in the dark without the need for any infrared cameras.
“The drone works by employing a software-defined radio and broadcasting spoofed GPS signals by making use of the ephemeris data of GNSS (Global Navigation Satellite System) constellations. The target drone’s GPS sensor locks onto our fake radio station transmitting at a much higher power than the available satellite’s transmission power. Following this, the drone generates fake GPS packets by mathematically modeling the time differences at the receiver’s end,” explained IIT Madras BTech student Vasu Gupta.
This ingenious way of sending out spoof GPS signals eventually helps in hacking into the GPS feed of the target drone. Furthermore, Vasu stated that by using four of the above stated time differences (from the fake GPS packet generation to the receiving by the target drone) the GPS sensor calculates the target’s 3D position and calibrates the drone’s time to the fake clock. Therefore, using this the researchers are able to modify the latitude, longitude, time, and altitude of the rogue drone.
The drone uses an advanced KCF tracker
“We have tested this electronic countermeasure of ours against nearly all the civilian GPS receivers used by the UAV industry such as ublox, DJI inhouse GNSS and we have been able to take down the drones almost instantaneously (within 4-5 seconds),” Project Associate Rishabh Vashistha said.
Apart from AI motion tracking, the IIT-M researchers have also made use of an advanced KCF (Kernelized Correlation Filter) tracker to track objects after the drone has a lock on them. This system is stated to be more efficient than sonars and radars as they do not provide much information about the object other than its physical position. Also, the researchers state the existence of a failsafe in the algorithm that would guide the target drone to a safe and controlled landing at the very spot. This would eliminate any chances of collateral damage or accidents.
Here’s a look at the drone in action:
