Issue 41 Unmanned Systems Technology December/January 2022 PteroDynamics X-P4 l Sense & avoid l 4Front Robotics Cricket l Autonomous transport l NWFC-1500 fuel cell l DroneX report l OceanScout I Composites I DSEI 2021 report

8 Platform one Zurich-based Fotokite has developed a tethered UAV for first responders that uses machine learning for control (writes Nick Flaherty). The Sigma hexacopter UAV is a fully autonomous tethered system based around a Jetson graphics processing unit (GPU) in the ground station to provide the machine learning. The tether allows the UAV to be deployed from a box on a vehicle to up to a height of 45 m without the need for an operator. The machine learning algorithms on the Jetson unit control the position of the UAV for autonomous launching and landing, adjusting its position to provide a stable platform for an EO or IR camera. The Sigma has been operated in winds up to 45 kph/25 knots, and it can detect hazardous wind conditions and will automatically adjust its height to a safe level to keep on station without the need to monitor or pilot the system. The UAV is built from carbon fibre and weighs 1.2 kg, and can still operate if one of its motors fails. It has a back-up power system for up to 120 seconds of flight and redundant flight controls. It carries an RGB camera with a 5x digital zoom and a resolution of 720p/ 30 fps, or a thermal camera with a video resolution of 320 x 256/30 fps to locate warm objects such as missing people or animals. Both cameras are supplied by FLIR, which is now part of Teledyne. The data from them is fed down a cable that runs alongside the tether using Ethernet for video processing on the Jetson unit. This is fed to a WPA2-encrypted wi-fi access point that provides live video streaming to a tablet computer for emergency staff. An integrated 4G LTE modem in the ground station also streams the live video to a remote incident command unit. The ground station handles automatic pre-flight system checks, the back-up power system for control and recovery in power blackouts, and provides an alert if the tether breaks. There are also automatic maintenance alerts generated by the algorithms on the Jetson unit that monitor the performance of the UAV and the ground station. “The next area of focus for us will be to increase the situational awareness and decision-making power in an emergency situation,” said Chris McCall, CEO of Fotokite. This will be via software updates to the Jetson module, he said. Airborne vehicles Tether feeds AI control December/January 2022 | Unmanned Systems Technology Machine learning algorithms in the Sigma’s ground station allow autonomous launching and landing The next area of focus for us will be to increase the situational awareness and decision-making power in an emergency situation