Prototypes

The first prototype UAV is built on a Lite Machines 110 remote control helicopter. The choice of the LM-110 is an ideal platform for debugging both the autopilot and the command and control software. It has nearly indestructable, full flapping plastic rotor blades and can take nearly any abuse from "software bugs" or Pilot Error.

Unfortunately, the 0.061 cubic inch (1cc) engine (by Norvel, pictured on the left) only produces about 0.1 horsepower. This is no where near enough power to lift the airframe, engine, embedded computer, servo controller, attitude gyros and camera. However, the larger aircrafts are very expensive to crash, so they are cost prohibitive for testing. The Enya-60 0.60 cubic inch (10cc) engine is on the right; it provides enough power for a 5 kg payload, but consumes far more fuel.

Instead of having all of the hardware onboard, the prototype has the servo controller connected to a laptop via a light serial cable. This tether limits the functionality of the UAV, but allows for easy debugging and off loads the underpowered LM-110. It can barely carry the camera, full fuel, servo controller and battery.

We now have new photographs of the first prototype with the camera head, the ground station and the video feed. On the right is the video feed test (click to zoom), showing the hotel room and television. The boom is still dented and the gearbox split, which you can see in the photograph at the top of the page (click to zoom in)).

The new prototype is in the process of being built. Some parts are on order, others are currently being modified. They are:

Kyosho Concept 60

The rotor diameter of the Concept 60 is over twice that of the LMH. This leads to over four times the blade area and lifting surface.

The Kyosho head is pictured on the right and the LiteMachines on the left. The simpler head of the LMH is due to its lack of any collective controls, while the more complex rotorcraft needs the ability to change the pitch of all blades at the same time in order to maintain a constant RPM. Here is an alternate image of the Kyosho head.

The height is also quite a bit taller. You can see the engine heatsink on the LMH quite clearly. The Kyosho has an active cooling system that uses an engine driven fan to force air over the head, but it is enclosed inside the mainframe.

The Hiller paddle and fly bar on the Kyosho are almost as large as the main rotor on the LMH. The Bell/Hiller subrotor on the LMH is clearly visible.

The engine, clutch, transmission and gearbox are all enclosed in tightly fitting plastic ducting. This means that it will be difficult to add onboard electrical power generation (such as the Genesys from Sullivan Products ). As you can see from the left and right side pictures (click to zoom in), there is very little room for the alternator disk and pickup. It's not even clear where to mount the HAL effect sensors for the N1 and N2 tachs.

Atmel AVR

The realtime tasks that it will handle are:

• Servo sampling (for the safety pilot receiver) [Done]
• Gyro sampling (and rate integration for the AHRS) [Done]
• Read engine and control sensors (tach, etc)
• Read NMEA data from the GPS
• Output PWM signals for the control surface servos [Done]
• Output PWM signals for the payload controls [Done]
• TTL interface to the payload, if necessary
• Write summary data to the nav/autopilot system on the serial port. [Sort of done]
• Read commands from the nav/autopilot system from the serial port. [Sort of done]

Suspicious.org's embedded 486

AHRS

Another option is to purchase three piezo gyros and use GPIO lines to read the results. It would load the system a bit more to poll the accelerometers, but the cost would be in the $300 range. This option was tried and discarded in favor of building a home brew IMU with gyros, accelerometers and a microcontroller to perform the rate integration. It is visible on the left, partially populated with one dual axis gyro and one dual axis accelerometer (click to zoom in).

You can see a screen shot from the curves program from the 1.5 release. It samples the gyros and performs rate integration on board. The results are then sent to the ground station for display with the EFIS::AI widget.

The MIT project used the Crossbow IMU, but the single unit prices are in the $4,000 to $6,000 range. I've written a survey of the available units.

The first prototype was built in September 2001 and flown under manual control via the computer in October. It consisted of:

Lite Machines 110

Mini-Servo Controller

Logitech Wingman

X-10 wireless camera