antenna-theory.com

[nivm321]

I am thinking of making a system which can generate a antenna for a given input radiation pattern, that is, we draw a radiation pattern we desire and the system churns out a antenna design which can achieve that? Is it trivial or is it valuable?
I have an idea of making a antenna design system but I don't know what input it should take. Any pointers?

Thanks

[hagster]

[nivm321]

How is it generally done? Say I have a radiation pattern I want to achieve, how should I go about designing an antenna for that? Trial and error?

And I will have to stick to only one kind of antenna, say maybe wire antennas. So maybe not very ambitious?

Thanks

[hagster]

I am not aware of a general solution.

Antenna engineers usually rely on there knowledge of prior art as a starting point and build/test/simulate from there.

There is however much more to antenna design that achieving a given beam pattern. Things like size and bandwidth are also important. Not to mention fit/form and construction etc.

My advice would be to do some research on current electromagnetic simulation software.

Some instantiations of NEC2 do can use genetic algorithms to adapt antenna designs to fit some fitness function. You could extend this functionality to try and fit a given radiation pattern. NEC2 has some limitayions on the types of antenna you can model, but it does run fairly quickly which makes it a feasable to use with itterative type solutions such as GAs.

OpenEMS is another open source tool that I beleive offers more capability than NEC2. I havent had a chance to try this properly yet though. A really great project IMO would be to create a descent GUI for this.

Commercial tools can probably do something similar to the GA of NEC2, but these tools are out of my price range.