The product and market needs
The technological challenge of the EXO Data project is to develop a flexible solution to remotely and in real time control the Exodronics fixed-wing aircrafts by means of 3G-4G LTE telecommunication network (ready to be adapted to 5G). The innovation will provide high control range beyond the line of sight of the pilot (BVLOS), accordingly with its flight autonomy superior to those currently offered by the market, allowing the integration of computer vision technologies, big data, deep learning, adapted to the new autonomous drone fleet paradigm.
The initial and main requirement was the unlimited range of communication between the drone and the ground segment, accounting for telemetry and live video. This requirement involves overpassing the local radio-link communications between the Ground Control Station (GCS) and the drone Flight Controller (FC). The objective is focused by using the 4G network, which allows to link the drone with the GCS and with any device connected to internet wherever it is located (GCS laptop, EXO Data server, operator’s headquarters computer, final user’s mobile device…).
Fig. 1. Example of EXO Data logical schema. Links description. A: FPV, telemetry and HD video (optional). B: FPV (optional), telemetry and HD video. C: CV processed HD video, telemetry insights and Data Base storage. D: Telemetry and FPV requests (optional).
Hardware and system architecture
The hardware and system architecture have been developed in collaboration with UAVMatrix [https://uavmatrix.com/] and NEXIONA [https://nexiona.com/]. Inside the drone there is typically a set of navigation sensors (GNSS, IMU, barometer, compass, pitot tube,…) that feed the Flight Controller (FC) [http://www.proficnc.com/], and a telemetry module with an antenna to radio-link the FC with the Ground Control Station (GCS). In EXO Data we need an extra microcomputer (e.g. Raspberry Pi [https://www.raspberrypi.org/products/raspberry-pi-zero-w/]), a modem (e.g. USB-stick) and a SIM card to connect the drone to internet. It is worth noting that the extra hardware is powered and regulated with an independent power circuit from the drone battery and power module, to avoid the supply thorough the FC and to keep the independent functions.
The video cameras must be compatible with the microcomputer ports. In the case of Raspberry Pi, we use a ZeroCam for First Person View (FPV) video, and for downwards Earth Observation video cameras we use a PiCam or a GoPro like camera (with a HDMI -CSI converter) connected to the Raspberry video input port.
Fig 2: Logical block diagram of drone elements and connections. Links description. E: FPV video (medium bandwidth). F: Downwards HD video (high bandwidth). G: Telemetry control (low bandwidth). H/I: Telemetry data (low bandwidth). J: Control commands (low bandwidth). K: Local bidirectional configuration/testing link. L: Local bidirectional configuration/testing link + FPV video (medium bandwidth) + Downwards HD video (high bandwidth) + Telemetry data (low bandwidth)
The system architecture is based in a Virtual Private Network (VPN) composed by three principal IP address: 1/ The on-board drone location. 2/ The GCS pilot location. 3/ The EXO Data server location. Each of these three VPN endpoints have access to internet and are securely linked between them, with an internal fixed identifier.
Fig. 3. Virtual Private Network (VPN) architecture schema.
A software is installed in an on-board microcomputer, which gets the telemetry from the Flight Controller (FC), the video from a camera, and is accessible to the VPN endpoints for its visualization such a web interface. Also, it can stream both telemetry and video data to the VPN endpoints and reproduce the data locally in the endpoint. The microcomputer acts both as a hub between the drone input data to be sent (FC and cameras), and as a bridge between the drone and the VPN (e.g. ZeroTier [https://www.zerotier.com/]).
Regarding the Real Time Streaming Protocol (RTSP), we use TCP/IP protocol to securely transmit the telemetry without losing data packages, while we use UDP protocol to transmit the video and assure a combined low latency. The telemetry data encoding and packaging is done following the MAVlink protocol, which can be decoded and plotted in widely used in GCS software (e.g. Mission Planner [https://ardupilot.org/planner/]). The video encoding and packaging is done with H264 compression (1:125 ratio), accounting for a video quality of 1280×720 pixels, 720p and 30 fps, which can be decoded and reconstructed in widely used in video software (e.g. VLC, Gstreamer).
Fig. 4. Practical case: hardware configuration and connection schema.
Is there 4G coverage when the drone flies?
The major concern is the 4G network telecommunication feasibility, specially the link cuts due to the coverage lost. In collaboration with EACOM [https://www.eacomsa.com/], we carried out theorical and practical experiments to demonstrate it.
Typical antennas deployed for the 4G network have vertical radiation patterns with beam widths that are around 10°. The antennas are subjected to both mechanical and electrical inclinations (tilt) to get them to focus on the coverage area and avoid overreaching with neighboring stations.
The following graph makes an estimate of the maximum and minimum distance considering a flight height of 60 meters and 120 meters and entering the station with an angle of 20° and another of 5° on the horizontal:
Fig. 5. 4G antenna beam lobule distribution pattern: theorical connectivity range from 4G antennas. (Source: modified from Kathrein).
The loss in free space (ITU-R p.525.2) is:
Lbf = 32.4 + 20 log f + 20 log d
f = frequency in MHz; d = distance in Km
We will take for the calculation the most restrictive one that is the 2600 MHz and the greater distance (1376 m):
Lbf = 32.4 + 68.3 + 2.27 = 102.97 dB
Signal level at the drone antenna = 43 dBm + 18.5 – 102.97 – 18 = -59.79 dBm
If the displayed frequencies were lower, the signal level would improve. A good signal level in 4G is considered above -75 dBm.
Therefore, it is expected that the drone can receive a good quality 4G signal at distances greater than 2 km from the station or even greater. However, the movement of the drone will cause at some point to enter a zero radiation from the station (see radiation diagram). This will force the start of a handover protocol to transfer communication to another node on the network.
In practice, we found that at flying heights under 120 m the coverage was better than at ground level. The topography and the surface elements (vegetation, buildings) are not shadowing the signal when the drone is in the air, while it was compromised before taking off. Moreover, if the 4G network antenna covering the flight area is located in a hill over the drone top altitude (as usual), the main antenna lobule is covering the drone 4G receiver and the signal is excellent. Nevertheless, if the 4G network antenna is located in a basin position (not common), or the drone if flying at higher altitudes that 120 m above ground level (commonly not allowed), the coverage quality drops.
Is the bandwidth enough?
Regarding the throughput of the network, the 4G network is not symmetrical, having less throughput in the upstream channel than in the downstream channel. In the case of streaming video transmission, the limiting factor is the throughput in the upstream channel, from the drone to the ground points. The standard specifies a maximum of 100Mbps in the downstream channel and 50Mbps in the upstream channel. However, being transmission systems that share the medium with several users, the actual throughput that we are going to obtain will depend on factors such as the load of the network at that time. In the theory an upstream throughput of around 9Mbps is to be expected, which may be sufficient for FPV. In the practice, we found a range from very low rates to peaks of 15Mbps.
Is the data consumption so high?
Taking as an example a 720p video at 30fps, the calculation of the storage needs in MB for one minute of video can be done as follows:
C = (1280x720x30 x (No. of colors / pixel) x (No. of bits / color) x 60 seconds x codec compression rate) / (8bits / Byte)
Assuming 3 colors per pixel and 8 bits per color:
C = 5GB x codec compression rate
One minute of video will require approximately 40MB of memory. We consider that a bandwidth of at least 5Mbps will be necessary to transmit this video in streaming with the appropriate quality.
Taking as an example a video in 4K quality and 60fps, one minute of video will require approximately 400MB of memory. We consider that a bandwidth of at least 25 Mbps will be necessary to transmit this video in streaming. This is a really high consumption of data, which makes it no affordable for common users, but with the upcoming 5G technology this scenario can change.
Is the latency low enough?
Latency can be the most critical point, especially in first person view (FPV) video, which must reach 250 ms. Depending on the position of the VPN root servers, the latency can reach 100 ms only due to the system architecture. This added value to the latency of the electronics (between 30 ms and 50 ms) and the one introduced by the 4G network, can in some cases cause problems for the video in real time (UDP protocol). In practice, it appears that although there are small micro-cuts, the quality of the FPV is good at a resolution of 720p and 30 fps and the latency is under the 250 ms requirement. However, this value is very volatile and dependent on the signal quality.
Multioperator SIM card?
There are some telecommunication companies that have agreements with the main telecommunication operators and provide “multioperator” SIM cards (e.g. Wirelesslogic [https://www.wirelesslogic.com/]). If a multioperator SIM card is in the drone on-board modem, it has the advantage to connect the operator network with better signal. In most of the developed countries, there are three or more physical network managers (e.g. Vodafone, Orange, Telefonica…) and they share the infrastructure, but everywhere there is better signal from one operator. The multioperator SIM analyses the coverage when initializes and gets operator with the best signal to perform this single flight.
Authors: Robert Guirado, Joan-Cristian Padró, Albert Zoroa (extra: EACOM)