Our primary airframe is a Swallow UAV, boasting a 2.8 metre wing span. Each upgrade in airframes has resulted in larger size, and this time is no different. While our previous Hugin airframe served faithfully for two years worth of competitions, even bringing us to second place in the national 2017 USC competition, we were pushing its limits in terms of payload capacity and flight time. This larger platform will allow us to increase the payload and promise greater endurance so we can be more competitive. The Swallow will use a Pixhawk autopilot board with modified code to meet strict safety requirements for competitions. The onboard computer is an Odroid XU4 which interfaces with our camera.
For the first time in the our history, Team Guardian is planning on flying two UAVs simultaneously. Our secondary system is a hexacopter, which will allow us to obtain greater detail on specific areas in competitions and drop or pickup objects with greater accuracy. A Pixhawk autopilot board will again be used to automate the flight of the hex. An FPV system will be used on the hex, allowing for manual control when out of range of sight.
Previous UAVs & Test Platforms
A Hugin airframe is our current test airframe. It was modified to hold our gear for two years of competitions. With a 2.2 metre wing span, this airframe was the largest that Team Guardian had employed while it served as our primary airframe. It's increased size allowed for greater payload capacity. The empty airframe weighs approximately 2.5 kg, but can take an additional 2.5 kg payload. The Hugin has approximately a 20 minute flight time with one battery. One of the greatest features over the previous airframe is the space available to us, almost 9000 square centimetres inside the fuselage to store our onboard equipment. Inside the Hugin was a Pixhawk autopilot board with modified code to meet stricter safety standards. The vision system used a Point Grey Chameleon 3 camera and an ODROID XU4 as the onboard computer, which is much faster than the Raspberry Pi 2 used in previous systems.
Our previous UAV was built on a Bormatec Maja, sponsored to us by Xiphos engineering, with a number of modifications that better suited our needs. Key considerations in our system are to keep it easy to disassemble and transport, as well as having easy access to the on-board avionics. The autopilot is an ArduPilot Mega, an open-source system based on Arduino using an Inertial Measurement Unit and GPS for telemetry and XBee's for wireless communications to the ground station. The payload vision system consists of a Canon G9 camera interfaced with a Raspberry Pi single board computer used for capturing sequential still images that are geo-referenced based on the aircraft's position and attitude. Images are transmitted to ground for post processing via Wi-Fi using all custom written software.
The UAV is designed to be small, portable, and inexpensive. The old UAV that we had was modeled on an E-Flite Apprentice trainer; it was modified so that it could be disassembled and stored in an equipment case. It was coupled with an open source hardware and software autopilot system, the ArduPilot Mega. The autopilot utilized an Inertial Measurement Unit (IMU), GPS attachments for telemetry data and an XBee wireless communication system. Video surveillance footage was relayed to the Ground Control Station via an off-the-shelf wireless video system.