DME University Project - Fully Autonomous Quadcopter for payload delivery


#1

I am a university student in my second year at the University of Surrey. We have been assigned a project as part of a second year module with the brief:

‘Design and construct a device capable of delivering a payload of water to a target area at
altitude, for use in a fire-suppression application. Submissions will be assessed based on
their performance in a simulated emergency situation.’

My team and I have come up with a solution fulfilling all the specifications provided with the brief of using a quadcopter. Due to the design having to be fully autonomous and a small budget of £250 we have been looking for a control system that will allow us to complete the mission without any human interaction from the initial take off to landing. Arducopter was something we found and I was wondering if this would have the appropriate qualities for the job?
Is there a particular accuracy the system as as well as we have to deliver the payload to a small target zone of [ 600mm x 400mm ]
The target zone is positioned 60m down a flight zone and we also have to have a method of releasing the payload, also autonomously. We require a system to guide the quadcopter to the target zone and deliver the payload. Any information or ideas would be much appreciated.


#2

Sounds like a fun project! If budget is the main thing and you dont mind fiddling around with loading firmware and editing some files yourself then I would suggest you go for either the multiwii based system with i2c Nav and GPS addons. However if you want something that runs the arducopter code (has much better features) then I would suggest you go for either APM2.6 system (easiest option), or the AIOP v2 board.

Overall if you have the budget I would go for the APM2.6 system as it comes with a better GPS module and is much easier to use and setup compared to AIOP.

Otherwise will be great if you keep us posted with your progress!


#3

Thankyou for the advice.
Yes indeed budget is the main issue we have as it is rather small; and that has to include the control system and components for the frame and contruction of the quadcopter itself.
We have begun to optimise the final design for the frame and will begin manufacturing it within the next couple of weeks or so. Is there much programming invovled with the APM 2.6?
For the payload release we are using a precision micro servo that will release a bung from the base of our cone payload tank when the quadcopter reaches the target area.
Due to the small size of the target zone we may have to use an LED sensor system to get the quadcopter accuratley above the target zone when dropping the payload. For the competition (in May 2014) more points are awarded for the more payload we get into the zone.


#4

Just a few clicks of the mouse to load code onto board. There is a wizard that walks you through everything including the initial calibrations.

As for the frame, if use wood you can make one quite cheaply and motors like the MT2213 are quite good along with 4in1 20A ESC. Otherwise something like the F450 frame is a nice option. You should be able to get it all done for around £100-120 excluding the flight controller and radio control.

The payload release should work well but if you have some spare money at the end it might be fun to investigate the electro permanent magnet idea


#5

That is very helpful as we were struggling with motor and prop configurations. But after some calculations I think we have got the right set up. We initially considered Balsa wood but unfortunatley don’t want to run the risk of it failing catastrpohically or be vunerable to fatigue. Instead we are going to use carbon fibre tubes and have these attached to aluminium plates using nylon clamps to save weight also. So far so good we are just modelling this in CAD and completing the FEA for the individual components and the vehicle itself as a whole.