|The University of Adelaide||Home | Departments | Search|
Autonomous Robotic Paraglider
Keywords: Unmanned Aerial Vehicle (UAV), Paragliding, Control, mechatronics
(Commenced: 01-Jan-2006,Concluded: 01-Nov-2006)
This honours project aimed to design and build of a fully autonomous robotic paraglider (see Figure) capable of negociating a predefined course within a circular workspace with a radius of 100 m. The paraglider is actuated using an electric motor driving an 11" propeller. Sensors include a GPS module and an altimeter. Inclusion of an Inertial Measurement Unit (IMU) and the development of an Inertial Navigation System (INS) were initially intended, but eventually proved to be outwith the scope of the current project. The paraglider is controlled by a microcontroller (Freescale 9S12 on a Wytec MiniDragon+ board). Measurement data is relayed back to a base station using a reliable radio link. The latter can also be used to fly the paraglider remotely using the GUI driven base station software.
The GUI driven control software is a very flexible program which allows the paraglider to be flown manually using a joystick and/or a conventional R/C transmitter; in addition, waypoints can be defined using the mouse, causing the paraglider to follow a predefined path (autonomous flight). Finally, in case of an emergency, the R/C transmitter can be switched on, taking over control from the microcontroller. The paraglider returns to autonomous mode when the transmitter is switched back off. This also allows the paraglider to be launched and/or landed by a human operator. A system overview is shown below.
The project set out by constructing and experimenting with a prototype parafoil. This helped gaining some initial experience with flying a paraglider. However, the prototype parafoil, made from inexpensive materials, turned out to be too small. As its construction is very time consuming, a commercially available parafoil was bought, capable of carrying a total mass of around 1.5 kg.
A gondola was designed which would allow all control systems to be tested on the ground (tricycle mode - steering is achieved by actuating the front wheel instead of the parafoil). In the end, the gondola and on-board electronics turned out to be too heavy for the chosen parafoil. However, a short flight was achieved in manual mode, after stripping the paraglider down to an essential constellation.
The radio communication system was developed primarily to log the system states and reconfigure the desired trajectory during autonomous flight. Various safety precautions were implemented, most significantly allowing a human pilot to take control of the system at any time by toggling a switch on a conventional radio transmitter. An overview of the software concept is shown below.
The dynamics of the system were modelled using the kinematic and dynamic equations of a tricycle. Control strategies were developed and tested using a MATLAB based simulated environment. The simulation environment also allowed testing of communications algorithms and source select hardware. The validity of the simulation was verified using real world data. The control strategy developed uses a proportional integral feedback loop for velocity control, and a simple proportional control to steer the system to a list of updateable waypoints. The controller successfully navigated the tricycle through a difficult trajectory, relying completely on position and velocity updates from a GPS module. The results of the real world autonomous control trial agreed well with the expected result obtained by the simulation.
In conclusion, despite the failure to achieve sustainable flight, the system is of sufficient flexibility to be used as educational tool and as a basis for further development. All software is available from this web page (see below).