Hi Rahul,
On 4/7/25 5:59 PM, rahul pillai wrote:
Hey Marcus,
Thanks for replying.
- I'm aware this is a big and ambitious project, but to me, most of the complexity lies in
developing the application window.
Ambitious is a good word here! Make sure your proposal makes clear which parts you're
going to do first, and which you can let slip off at the end, if you don't manage to
finish every single aspect of the project.
The signal processing library and OOT modules are definitely significant parts of the
project, but I already have solid experience developing
with beamforming and DoA estimation algorithms from my work.
That's good! Maybe upload an example of one of these algorithms implemented in C++ or
Python to your github account and link to that in your proposal!
I've also used GNU Radio extensively, especially with Embedded Python Blocks
for quick prototyping. So out of the three main components of the project, I’ve already
tackled two before, and I expect those parts to take
less time to complete. I haven't exactly developed OOT modules before, but all it takes is
a read through the documentation.
I'd think you might be right, if you know which algorithm to implement and have done some
of that before, it can be a reasonably manageable task.
Having advised a master's thesis on adaptive beamforming, however: I'd feel much more
comfortable if you limited yourself to a single one, preferably one that you've
implemented before, and would move the rest to "stretch goals".
It's better to build something slimmer that works, which someone (you? A student you'd
mentor later on?) can later extend with more algorithms, than to build something with
multiple competing algorithms, but then not finishing integration. Remember that you'll
not only have to program each algorithm, but prove – with tests you write – that they work
as well as expected.
- Timeline is one of those things I am yet to complete.
I'd predict low chances of acceptance to proposals with an own project idea without a good
planned schedule; I'd encourage you to prioritize that! (it goes hand in hand with
deciding which features you might want to make "optional".)
- Those 4 different visualizations will be part of the plotting lib of the application
window. And I am the one who made it, the code is available in
repo <https://github.com/Ashborn-SM/GSoC-Proposal/blob/main/plots.ipynb>, including the
one at the title page.
Suuuuper Cool!
Best regards,
Marcus
