Initial Phase
The initial design followed mainly the shape of a fish, utilizing mechanisms on the commercial airswimmer, in particular the toothed gear and track to achieve vertical mobility. A pair of cameras was envisioned to provide a wider field of vision or even peripheral vision. All the electronic components were to be mounted on the buoyancy adjuster, so as to increase the resultant tilt.
Prototype
Mock-up of the components
The picture above shows the early mock-up of our components. From the top, we can see the tail mechanism and the servo, which runs to the Raspberry pi. From the Raspberry Pi, the wire goes to the bladder mechanism servo that will move the battery holder. The battery holder contains the Li-po battery that we used as the bladder which also powers the Pi. Finally, the pi is connected to the camera that allows us to live-stream whatever the HoverFish sees.
Tail Mechanism & Touchscreen UI
The following video the early prototyping and proof of concept of moving the tail servo using Raspberry Pi. The UI is still in its infancy, but we can send an input from the device to the Raspberry Pi. In turn, the Raspberry Pi controls the servo angle to flaps the tail according to our input. The development of the UI and the coding can be found here, while the development of the tail mechanism can be found here.
Bladder Mechanism
At the time of prototyping the bladder mechanism, it was just a theorised way of playing with the centre of gravity and the centre of buoyancy. As we do not have the proper size mylar balloon at the time of prototyping, it is very hard to know whether this concept is practical (the development of the bladder mechanism can be found here). However, we have tried the bladder mechanism on a smaller balloon with considerably lower load on the servo, which proves that this method is applicable when we scale it up. The following video shows the change in tilt of the small balloon as we move the servo:
Video Streaming Quality & Movement Input
The following video shows what the HoverFish sees and the input we gave it. The camera resolution is 600 x 480 pixel @ 24 fps. Explanation on the UI can be seen here.
Putting All Together
After several attempts of getting used to the mylar and discovering various points to take note while working with it, we developed this design to reduce the number of corners to be heat and bonded. More detail of the balloon making process can be found here. The design also shows the location of various components on the balloon.
Final Product
Here is the link to us flying the HoverFish for the first time in front of the Year 1 Physics Lab:
Seeing that it can move and manoeuvre around, we play with the HoverFish around the SPMS MAS Atrium. Here is the full video: