However, this new drone is superior in at least a couple of ways. One, it's much more stable in flight, thanks to its 68.5" wingspan. Its size also means it can loft a larger payload. The photo above was taken using an 11Mpx GoPro Hero 2, which is small, but has a not insignifcant weight penalty.
JournoDrone 2 was a plastic shell that I wrapped in carbon-fiber and epoxy, which could take a crash on the nose without much harm. This newest drone is made of balsa wood, but it's such a docile aircraft that the need for crash resistance is minimal. All that balsa, some 5 or so pounds of it, is also pretty good at flexing and absorbing a hard landing.
As you can see, the camera setup was extremely simple: the GoPro was attached to the bottom of the drone using electrical tape. The camera was then programmed to take photos on a time lapse, at 2-second intervals. The camera angle created by this setup was not precise or ideal for aerial mapping, but it was enough to test the concept.
The drone didn't come in Illinois livery -- that was a custom job for the grant this drone is meant for. This particular drone, when it's in 100 percent working order, will go to a Central Illinois high school, and was funded by a National Science Foundation grant to improve K-12 STEM education. There, students will learn about the physics, computers, and robotics that make the drone fly. They'll also be putting it to use, flying missions and taking aerial photos of things of environmental and ecological importance near their community.
The system works great for rudimentary manual flights where precision is not required. But to ultimately make photo maps, or point a camera at a fixed position on the ground, a GPS-guided automatic pilot is necessary.
To this end, we're using the 3D Robotics ArduPilot Mega (APM) version 2.0. This microcontroller uses an on-board gyroscope, accelerometer, barometric pressure sensor, and GPS receiver to guide the drone. It also serves as a sort of black box, enabling us to figure out what happened during a mission.
For example, here's a flight log that has been turned into a KML file and rendered on Google Earth. I did notice in this flight data, however, a "ramp" approximately in the middle of the flight seemed to suggest the GPS was losing a satellite fix.
Fortunately, the autopilot also keeps track of when it has a satellite fix. Above is a graph that the mission planning software generated after looking at the flight data. A 1 on the y-axis indicates a fix, and 0 indicates no fix. The x-axis indicates the number of lines of data the autopilot has recorded, which is proportional to the mission time. The graph clearly shows that the autopilot lost a fix during the flight.
Subsequent flights show difficulty holding a satellite fix. This could be for a number of reasons: the autopilot isn't getting adequate voltage, the servos and/or esc are causing interference, the GPS isn't getting a strong enough signal from inside the avionics bay, or the GPS itself is defective. I'm likely going to replace the GPS with a module that has a larger antenna to resolve the problem. That will allow for more accurate photos and photomaps.
Here's a short teaser video of the drone in flight, which we made to increase excitement around the Drones for Schools program. If you don't like The Chemical Brothers, I would recommend muting the video:
Some additional aerial shots:
And finally, some quick specs on the drone:
Wingspan: 68.5 in (1740 mm)
Wing Area: 722 in² (46.6 dm²)
Weight: 7-8.5 lb (3175-3855 g)
Wing Loading: 22-27 oz/ft² (68-83 g/dm²)
Length: 56 in (1410 mm)
Motor: 42-50-800kv brushless .32
Primary Battery: 6-cell Lithium Polymer 2400mAh 22.2v
Radio Battery: 2-cell Lithium Polymer 640mAh 7.4v
ESC: 80 amps
BEC: 10 amps
Primary Camera: 11Mpx GoPro Hero 2HD 1080p
Secondary Camera: 5Mpx micro camera
