Built to play a role as a character in an outdoor theatre production, the Hummingbird was a fully flight capable animatronic Tri-copter drone. The first of it's kind in the UK and quite possibly the world.

The early R&D involved lots of ideas and building several versions of propeller driven multi-rotor in various configurations. Trying to find one that was stable in flight while still having something of a bird like character about it.

It was important for the character to actually look like a bird. so I modelled the form of a hummingbird from reference images, and scaled it to the size that would work with the electronics I needed to put in.

From this I was then able to develop a series of plates forming an internal chassis that everything was fitted to.

At the same time, I built two Quadcopters and a Tricopter so I could learn how to fly drones. With the ultimate aim of being capable of performing with the thing I was building.

During the build process I realised I would need to get some kind of license to fly it from the Civil Aviation Authority (CAA). One of the obstacles this brought was that I also needed to make it capable of flying and landing autonomously in case of a loss of signal from the controlling radio.

Inside the head of the hummingbird, I mounted a forward facing high resolution camera and a GPS satellite receiving antenna. And inside the body the internal brain of the bird was actually very smart. The final bird was not only capable of stabilizing itself in flight via a gyro and inertial measuring unit, but also of knowing where it was at any given moment in height via a tiny barometer and precisely on the earth via the GPS antenna in it’s head.

All of the sensors onboard the Hummingbird drone gave it the ability to find it's way back to it's take off position in the event of a battery low or major radio interference or complete signal loss. Of course I tested all of this, and below is a video of one of the tests where the drone returns from several hundred metres away. The video was quite long so I'm just showing the last bit here.

I also experimented with a short wave radio in the tail enabling it to communicate with a laptop on the ground. This meant it would have been able to fly autonomous missions programable via the laptop ground station and with the feedback from it’s internal guidance system.

The head was fully articulated and able to look around and up and down as he flew. And also once airborne, the landing gear (the rods with green tips) folded up and his tail was programed to fold down and then to articulate to and fro in sync with the flight movements themselves. This was to help with balance like a real bird does and also to give a better overall look and feeling of bird-ness.

I ended up experimenting with tethering the drone when performing with it. And after meeting the Civil Aviation Authority, showing them my tests and discussing the project with them they told me it was safe to fly it like this. The main advantage was that it meant the bird could fly much closer to the audience than otherwise possible under CAA regulations.