ThermoPilot, a Thermal Hunter Glider Drone
This project has been started in october 2010 and its purpose is to build a full autonomous Thermal Hunter Glider which is able to fly for a long duration in an area defined by a radius, a ceiling and a minimal altitude. The glider uses thermal soaring and dynamic soaring methods commonly used by full scale gliders during cross country soaring competitions.
During a full autonomous flight defined by a flight plan, the glider is able to optimize itself the onboard power using the energy gained by the updrafts during its thermal hunting process. For safety reasons, the glider can use its electrical engine to sustain the minimal altitude required.In the video, you will see some samples of the test flights during the phase 2 of the Thermopilot project. The glider is fully stabilized by the Thermopilot software.
During the Phase 2 of the project, the firmware has been tested in flight with an ArduPilot/ArduIMU board (firmware ThermoPilot v2.9.x JLN version). The ThermoPilot (phase 3) will use an ArduPilot MEGA 2560 and a bigger glider. The setup has been successfully tested on the MAJA drone in May 2011. The ThermoPilot will be able to learn and optimize itself the thermal upwards areas during its thermal hunting...
The ThermoPilot firmware has been tested in High Mountains in July 2011 (Savoie, French Alps): Ridge Soaring tests flights done in high wind conditions.
It has also been tested in the fields in August 2011 (Auvergne, France): Thermal Soaring tests flights in various weather conditions.
The Cularis UAV autopilot setup is:
- Autopilot: Ardupilot (Atmega 328)@16MHZ + ArduIMU flat v2
- ardupilot v2.9.1 (JLN) + arduimu v2.0.1 (JLN)
Sensors: (arduIMU on board):
•3 Axis Accelerometer (ADXL335)
•3 Axis Gyroscope (LPR530AL + LY530ALH)
•3 Axis magnetometer (HMC 5843)
•1 pressure sensor (air speed sensor)
•1 Mediatek MTK16 10 Hz GPS
•1 GoPro HD Wide (forward view)
More infos at: http://diydrones.com/profile/JeanLouisNaudin
Here is a link to my ardusoar project with radian pro and apm 2.6 automated thermaling
Here some news about the ThermoPilot v5 project currently under developpement (now in phase 4). The Thermopilot algorithms have been now successfully moved to the latest version of the ArduPlane v2.28 firmware (official release) wich is now the named ArduPlane v2.28ThP5 version for the project.
The ThermoPilot project have been started in october 2010 and lot of real tests flighst have been already done in summer/spring 2011 in fields and in also high mountains. I have started the project with a simple ArduPilot IMU board on a Cularis 2m60 electro-glider (phase 1, 2) and then on a ASW-24 3m50 electro-glider with an ArduPilot Mega 2560 v1.4 IMU (phase 3). The phase 4, now uses a APM v2 with the v2.28ThP5 firmware. My new firmware version uses a machine learning algorithm (ML) to do an accurate thermal hunting during a full autonomous cross country soaring mission of the glider.
I am using Xplane v9.70 as a test plateform in HIL mode because, it is able to reproduce with accuracy various weather conditions (thermal upwards, dynamic soaring, slope and/or ridge soaring) in real mapped world and fully compatible with the google map environnement...
Below some screens about the v2.28ThP5 tested in HIL simulations on Xplane v9.70, this new release of the firmware will be tested on the ASW-24 as soon as the weather will be less cold and better for the thermal hunting...
Below the flight plan used as a cross crountry soaring mission:
Along the predefined track, if the glider found thermal upwards, it uses them to gain altitude and then continu following the route.
In HIL simulation the glider can fly indefinitely without the need of the engine... Its really fun...
More infos about the recents tests and dev HERE
Very nice project.
I'm interested to know if you've been able to maintain dual aileron and dual flap control while using the APM. Can you get it to mix the 4 channels normally required for glider ailerons and flaps or do you just use a single channel for each? I would also definitely be interested in taking a look at your code if its open and any presentations or documentation you may have on your project.
is the code open?
I started something like that and was developing when I crashed my plane, I too was using only gps altitude at 10hz and it was very good despite all the things people say about vario resolution.
I wish a code to start with, is yours available?
Very, very well done and my congratulations on your recent work and video just posted.
Here a video of the ThermoPilot v5 in action during a full autonomous Thermal Hunting flight (HIL mode with the APM Planner) with an ArduPilot Mega 2560. When the pilot activate the "THERMAL Mode" with the transmitter, a 3D reference point is recorded, the glider drone begins to do a full autonomous searching of the thermal upwards in the selected zone. The area is defined by a max radius from the reference point, a ceiling and a floor altitude. The main objective of the glider drone is to fly itself in this area and hunt the Thermal upwards to gain altitude...
Here a video of the ThermoPilot v5 in action during a Ridge Soaring flight (HIL test) on the APM Planner with the APM2560. Two turns points have been recorded through the transmitter by the pilot placed on the hill. Then, the glider do a full autonomous flight between these two points with the main objective: keep the altitude above these two points by exploiting the wind gradient and the dynamic upwards.
Here a video of the ThermoPilot v5 in action during a cross country soaring HIL test on the APM Planner with the APM2560. A flight plan (FPL) composed of 3 waypoints has been set in the APM2560 and the glider runs the FPL with the objective to hunt thermal upwards so as to keep the required altitude between the wp.
The firmware used here, Thermopilot v5.0 (JLN), is an updated version of the APM firmware v2.23....
The ThermoPilot project Phase 3 is now started...
The ArduPilot Mega 2560 with the ThermoPilot firmware v4.2 is fully installed and operational in my big ASW24 glider.
See some photos below:
Stay tuned. Regards,
Thanks for your comments. The Thermopilot (v4) version for the APM2560 will be able to build up (in theory, today...) an automatic mapping of the thermal updrafts area, this is that I have called the self-learning phase and it will be done during the thermal search mode. I need free memory for doing a such thing, this why I have used the APM2560.