A decade ago, drones were considered a technology restricted only to official authorities, such as the military, police, etc. But, numerous companies and civilians use UAVs (unmanned aerial vehicles) today for recreational photography, inspection, delivering goods and services. The Federal Aviation Administration (FAA) estimates more than 2 million drones are in use across the U.S. as of 2019. Of those, around 1.3 million are owned for recreational use.
This high population density of drones and frequent drone-related incidents pose significant safety and security challenges to critical infrastructures (CIs), especially around airport facilities.
Between 19-21 December 2018, hundreds of flights were canceled at Gatwick Airport, London, following reports of drone sightings close to the runway. It is estimated that 140,000 passengers were affected, with around 1000 flights either diverted or canceled, costing millions of pounds for the airport and airlines.
In a similar event, Newark Airport in New Jersey was closed due to a drone spotted in the vicinity for 90 min in January 2019. The incident caused USD 90M of economic loss.
Despite the FAA’s efforts to contain the risks of unsafe and non-compliant drone operations, the problem seems to be accelerating, with more than 2000 near miss sightings per year being reported by airplane pilots, air traffic controllers, and other aviation stakeholders.
This has created the need to detect and disable rogue drones and unauthorized drone activities, creating a new area of research and development in counter-drone technologies (C-UAS).
Countering a drone is a complex, multi-step process, involving interaction between several distinct sensors, methodologies, and communication with human operators. This includes three main categories: (i) detection, (ii) prevention, and (iii) mitigation. To detect rogue drones, airports use four types of sensors, namely radar, radio-frequency detection sensors, acoustic sensors, and visual sensors. Let’s look at the pros and cons of these commercially available drone detection systems used to defend airports.
A surveillance radar with single or multiple antennas sends out a signal to receive aircrafts’ reflection, measuring spatial coordinates and, optionally, velocity, acceleration, and direction.
- Long-range primary surveillance detection system up to 100 km, depending on RCS and altitude.
- Can track most drone types, regardless of autonomous flight.
- When combined with machine learning algorithms, it can distinguish birds from drones.
- High-accuracy tracking while in an angle range of observation.
- Able to track multiple targets simultaneously when using multitracking coverage.
- Bistatic and multi-static radars increase the accuracy of UAV detection.
- Independent of visual conditions (day, night, overcast weather, etc.)
- No need for RF or acoustic signal
- Detection range dependent on drone size and radar cross-section (RCS).
- Radar systems designed for manned aviation cannot detect small flying objects.
- High acquisition and installation cost
- Requires a transmission license and frequency check to prevent interference with other RF transmissions.
- Hard to detect low-altitude-flying, slow-moving, or hovering UAVs.
- No pilot tracking capability or ground control geolocation
- Lack of automation and high dependence on trained radar operators
- False positives with similarly shaped objects (birds, clouds, etc.)
2. Radio-Frequency Detection
Radio-frequency (RF) scanners use passive detection technology. This cost-effective solution detects and tracks UAVs based on their communication signature. They explore algorithms to scan known radio frequencies and find and geolocate RF-emitting drones, despite weather and day/night conditions.
- Lower cost than radar sensors with a medium-range up to 600 m
- Detects certain radio frequency bands where UAVs and GCS communicate for command and control (C2)
- Can capture RF emitted by UAVs and can locate UAVs and controllers
- Can capture WiFi-emitting drones
- High-accuracy detection
- Early warning capability even before UAV takes-off (when turned on)
- Triangulation is possible with multiple RF sensors.
- Machine learning algorithms can classify drone transmissions.
- Passive detection, no license required.
- RF signal cannot detect autonomous flying drones.
- Electromagnetic interference and loss of sight degrade detection capabilities.
- Variable detection accuracy depending on drone type and frequency band
- The attacker can spoof MAC addresses.
- Can detect only a few UAVs at a time
- Less effective in heavy-RF environments with a range of less than 100 m
- Detection limitations for swarm of drones
- Some passive systems may emit RF signals, despite being characterized as passive systems.
3. Acoustic Detection
Drone propellers transmit an audio pattern that can be detected and used for drone positioning and classification by acoustic sensors. Usually, a microphone detects the sound made by a drone and calculates the location using the time difference of arrival (TDOA) technique.
- Classification based on the acoustic signature
- Can differentiate between authorized and unauthorized UAS
- No need for RF signal for detection. Can detect autonomous flying UAVs
- UAV detection can extend beyond the line of sight
- Classification based on UAVs’ acoustic signatures
- Time difference of arrival (TDOA) technique is used for UAV localization while triangulation is possible with an array of distributed sensors
- Low-cost sensors
- Can provide drone direction or rough estimation
- Depends on an available library of already-captured sound signatures
- Higher false positives due to the increasing number of drone models
- Unreliable detection at range >300 m
- Does not work as well in noisy environments
- Detection limitations for swarms of drones
- Detection performance is affected by wind direction, temperature, line of sight, and signal reflections due to obstacles.
- They are not used as a primary detection source.
- No pilot tracking capability or ground control geolocation
4. Visual Detection
Electro-optical sensors in imaging systems and cameras use a visual signature to detect UAS, while infrared sensors use a heat signature. When combined with optical data, neural networks and deep learning algorithms can provide significant support and advanced intelligence to a UAV detection system.
- Need for human interference or artificial intelligence to efficiently detect UAVs
- Not used as a primary detection source (both EO and IR cameras)
- Both have detection limitations based on resolution capabilities.
- Hard to capture swarms of drones.
- IR and EO cameras need a direct line of sight to detect UAVs.
- EO Cameras depend on daylight and outdoor illuminance conditions (overcast, darkness, etc.)
- May confuse UAV with a bird or similarly shaped small airplane.
- Range limitations depending on weather conditions (clouds, rain, fog, mist, etc.)
There are several technological solutions for mitigating threats from malicious UAS when approaching critical infrastructures. Two types of C-UAS technologies exist: electronic and kinetic. Electronic countermeasures can defeat UAVs by using communications link manipulation, RF jamming, or GPS spoofing. Kinetic interdiction refers to intercepting UAS by physical means. We shall look at these C-UAS technologies in our next post.