Should i do build real model during Master's degree ?

[ammjy]

Good day, EEVBlog users.

Thanks you for your time.

Just i started my Master's degree in previous year.

and my thesis's theme is Attitude control of Multi Copter. and i choose Quad Copter.

But, still i am not sure.

Some peoples said, actually i don't need build a model. Even if that is a just only mathematical model, build model will be quite big work.
and actually i can take good model that included mathematical model in Matlab from other student in same lab. he's area is hardware and he also doing multicopter.

in other hand, Other some peoples said, should build my model myself to do attitude control.

i'm alone actually. i wanted carry out some project with peoples, but. my lab is small. so, i should choice what i can do myself alone.

Please, give to me advise...

[vealmike]

You need to ask your professors for advice.

Personally I'd say that you do need to prove your MATLAB models work by a physical implementation & demonstration. I wouldn't build a quadcopter, I'd buy one, rip out the electronics and replace with your own.

[ammjy]

Thanks you for your time and your answer, Mr. Vealmike !

Yes. you are right. i should ask to my supervisor. but, in recent time, he looks busy so i couldn't..

and actually your opinion is close with my think. Thanks you.

[Yansi]

Yeah, typical schools.  They give you work, they don't say you what to do with it or about it and you can't ask, because the teacher is so busy, you can't even find him (while he is doing close to nothing, sipping his coffee in the kitchen and chatting with his colleagues). But sury, busy as hell.

I am quite sick of how most of the universities are run.  And it seems it is not just a localized problem round here, eh?

[Rerouter]

Depends a little on the complexity of your attitude control, but I would recommend get a cheapy with DC motors, set it up suspended in a open frame wooden box, and play with that, that way you can rotate the box and see how it responds etc, with your own controller either on or off the device.

You add a single string with some slack and 2 knots through the center of mass, by limiting the slack it can move in the box so you can see how it responds, but its constained and cannot hit anything so long as you don't stick your hand in the box.

Equally the string gives you somewhere to run an umbilical if you want a more powerful device to talk to it. (String could be fishing line)

[brabus]

Everyone loves to see something ACTUALLY working, flying and doing stuff, even if not perfect.
Me too: I built a pathetic signal generator, but it was nicely done, it worked and put a smile on the board's committee. Everyone looked interested and came up with interesting questions - quite a moment indeed!

No one is excited about a bunch of MATLAB simulations and a boring presentation.


So... Go for it!

Pics or it didn't happen!

[ammjy]

Thanks you for your time, Mr. Yansi !

Yes, some peoples don't have interest to students. i agree with you strongly !

But, in recent time, my supervisor was really busy )) i think i wrote with possibility of misunderstanding.

Thanks you for comment !

[ammjy]

Thanks you for your opinion, Mr. Rerouter !
(I like your Nickname ! )

Thanks you. i got some idea from you.

might be i can minimize(optimization) my model depending on the situation as you said to me ! )

[ammjy]

Thanks you for your comment, Mr. Brabus !

Yes, i totally agree with you.

i have similar experience before, and actually that worked.

Always thanks you to EEVBlog users. giving to me many advice.

[Rerouter]

Just to clarify. Something like this. My idea was vertical so you could check if you had any yaw issues. But pretty similar. You can test it over and over again without having to worry about it flying into a fence at top speed.

[ammjy]

Thanks you for video and advise, Mr. Rerouter

gave to me lot of sense !