DeepSig Blog

Reinventing Wireless with Deep Learning

Introduction to Commercial Drone Signals

Posted by DeepSig on Aug 14, 2020 10:22:20 AM

Introduction to Commercial Drone Signals

Commercial and hobbyist drones come in many different shapes and sizes, both physically and with respect to their wireless control and data signal types. Nikola Tesla had one of the first visions of unmanned aerial vehicles and described his vision in the patent, “Method of and apparatus for controlling mechanism of moving vessels or vehicles,” that was granted November 8, 1898. In this patent, he suggested that these vehicles would be controlled by “waves, impulses, or radiations”.

Nikola Teslas patent on unmanned vehicles

Figure 1. Image from Nikola Tesla's patent on unmanned vehicles

Commercial drones are controlled and send data, primarily video, back over wireless connections. The types of wireless connections that the drones use can vary. Some use proprietary protocols while others use more off-the-shelf mechanisms.

Many commercial drones, such as those from early versions of DJI and Parrot, used 802.11 or WiFi, in the 2.4 GHz frequency band, as the technology for transmitted data. Many inexpensive commercial drones still use WiFi for control today. WiFi is convenient for manufacturers as commercial chips are cheap, readily available and allow consumers to control and receive data easily from their smartphone.

However, WiFi does have disadvantages. It is inherently limited in range and, when flying in areas with other WiFi networks, the signal quality may be degraded. This decreases the range even further or causes poor video quality. In recent years, manufacturers have been adding in their own transmission systems that provide for better interference mitigation and longer ranges. This complicates the development for creating systems that detect and classify different types of drone signals as the drone wireless communication systems ecosystem has become much more diverse each day.

WiFi-based Drone Wireless Communication Systems

drone wireless communication systemsControl and data packets are embedded in standard 802.11 frames on either the 2.4 GHz or 5.8 GHz band. In many cases the drones will automatically switch between the 2.4 GHz and 5.8 GHz bands based on the current wireless environment to avoid interference.

Common drones that use this transmission system:

Parrot Bebop 1, Parrot Bebop 2, DBPower UDI, DBPower Discovery, DJI Tello, Tenergy TDR, Wingsland, DJI Spark (without the DJI controller), Mavic Air (without the DJI controller)

Wi-Fi-based systems can easily be detected using standard 802.11 discovery mechanisms. The manufacturers are allotted blocks of MAC addresses which allow Wi-Fi survey systems, such as Kismet, to quickly discern whether a Wi-Fi enabled drone is in the local area. Most also have well-known wireless network identifiers (SSID’s) that can be detected.

Detecting and classifying signals for drones that do not use Wi-Fi for their control and data is traditionally much more difficult as many of the signal types are proprietary and can operate on a wide range of frequencies.

Non-WiFi Transmission Systems for Drones

There are thousands of drone companies that exist today with products on the market. Some are Wi-Fi and many are not. As an example, DJI has three separate proprietary transmission systems for their drone products alone. These protocols are less likely to be affected by interference and have a longer transmission range.

Common drones that use DJI’s Ocusync and Lightbridge (LB / LB2) protocols:

DJI Mavic Pro (Ocusync), Phantom 4 Pro V2.0 (Ocusync), Phantom 4 Pro (LB), Phantom 4 Advanced (LB), Inspire 2 (LB2), Matrice 200 Series (LB2) and Matrice 600 Pro (LB2)

The hobbyist / drone sport market also has a number of other general purpose controllers that use the ISM bands (915 MHz, 2.4 GHz, and the 5.8 GHz bands) with complex frequency-hopping controllers such as those made by FrSky and FlySky. There are also long-range frequency-hopping telemetry systems that support the MAVLink protocol which operate in the 433 or 915 MHz ISM bands.

For these systems, as opposed to WiFi-based systems, the uplink, or the control signal from the controller to the drone is a completely different signal type than the video feed coming from the drone to the operator. It is possible that hobbyist drones can in fact use frequency-hopping controllers, MAVLink telemetry, and WiFi video all on the same platform.

DragonLink Drone Range Extender-2In terms of signal characteristics, they range from very wideband OFDM signals, that typically carry data from the drone down to the controller, to small narrowband bursts that hop around the spectrum. These latter hopping transmissions are typically the control signals which need to be more robust to interference so as not to create a situation where the drone is not controllable.

Range extension systems, like the one shown in Figure 2, are also available for commercial drones that commonly operate at 433 MHz. This lower frequency provides a much greater range but does not allow enough bandwidth to provide a high-quality video link. There are many different systems that operate in the 433 MHz band, each having their own signal type. These signal types are typically some variant of a narrowband FSK signal.

Common range extension systems:

DragonLink, EzUHF, OpenLRS

OmniSIG SDK for Commercial Drone Signal Detection

The OmniSIG sensor and OmniSIG SDK are well suited for building systems that quickly detect and classify commercial drone signals across all of these areas. Whether they are 802.11-based and OmniSIG is paired with a WiFi post-processor or they are not 802.11-based and the OmniSIG SDK is used to train a neural network to recognize the signals, OmniSIG provides a innovative way to detect and classify commercial drone signals at low SNRs and OmniSIG SDK provides a way to add new drone signals to the system within hours.

Download our guide, Using OmniSIG SDK to Create a Drone Detection Model, to learn more about  SDK. In the guide, we provide an detailed example of using OmniSIG SDK to build an RF dataset that consists of DJI Ocusync uplink and downlink signals. We show how this dataset is used by OmniSIG SDK to train a new OmniSIG model that can be used by the runtime OmniSIG sensor to immediately detect and classify DJI Ocusync signals.

Use OmniSIG SDK to detect and classify commercial drone signals - use machine learning to detect commercial drone signals

Topics: Deep Learning, Drone Detection, Wireless Signal Detection