side lobe supression of antenna radiation pattern

I want to know how to supress the side lobe of the radiation pattern generated by the any type of antenna, without any information loss. I will be obliged for any help.

[Walter Roberson]

Can that even be done?? Without changing the antenna or putting some kind of enclosure around the antenna??

I do not think it is _possible_ to do what you are asking -- not unless your "without any information loss" is a question about applying error correcting codes to increase reception reliability after you have blocked part of the signal.

Or is the question one of how to design a reflection and focusing device around an arbitrary antenna pattern that will redirect the side-lobe energy towards the front (and back, considering that omni antennas radiate 360 degrees and the signal towards the back is manifestly not *side lobe* signal.)

[Honglei Chen]

Tapering can help reducing the sidelobe but I am also curious about what you mean by "without any information loss".

[Walter Roberson]

Thanks, Honglei, I had not heard of tapering before. I found a paper that goes through it with respect to stacked Yagi antennas, http://www.ncjweb.com/Ant_Interact_Part_4.pdf

It appears to me that the technique cannot be used with antenna of arbitrary shape: it appears to me that the technique requires multiple antenna segments whose position (and possibly orientation) to each other can be adjusted, and requires that the currents (power) to each segment be adjustable.

It is possible that the original poster is asking: given multiple antenna each of which has the same arbitrary beam pattern, how can one position the antennae and use current tapering to eliminate (or at least substantially reduce) the side lobes?

I gather (perhaps incorrectly) from the Yagi paper that it is not considered feasible to derive the positionings purely from theory, but that any given arrangement can be assessed by modelling the field interference, and that one can work towards a "best practical" arrangement by intelligent application of wave interference principles (e.g., model in position differences that differ by multiples of 1/4 wavelength.)

Complete sidelobe suppression did not appear to be possible in the stacked Yagi paper -- just considerable reduction.

[Gupteswar]

thanks, roberson and hanglei for your precious suggestions and guidelines. about information loss i will make the point clear. when the antenna radiates, the main lobe and side lobe contain the same information only. so supression the side lobe will not cause any information loss. rather it will increase the probability of secure data transmission. so please assist me if there is some way to code this in matlab.

[Walter Roberson]

You need to clarify: how many antenna are you working with, how much freedom do you have in their placement, list what you *are* allowed to vary, and list what you are specifically *not* allowed to do; and indicate if there are any other specific requirements (e.g., reduction of the side lobes by 50%?)

[Gupteswar]

yes sure,

i need to write a matlab code which can take antenna radiation pattern as input and at output it will produce the side lobe supressed pattern. the side lobes should be supressed 50%. and the parameter to be varied is the "array factor". and we are not allowed to manupulate "antenna dimention".

[Walter Roberson]

Those requirements are the same as being able to completely eliminate the side lobes for a set of identical but arbitrary antenna, placed in arbitrary locations, by merely multiplying the outgoing signal to each antenna by per-antenna complex weights.

As a single antenna is an example of the above placement criteria, this requires being able to completely eliminate the side-lobes for an arbitrary antenna radiation pattern, by multiplying the signal by a (non-zero) complex value. Sounds dubious to me.

Why do I say "completely eliminate" the side-lobes? Because if you give me an input configuration and require that I reduce the sidelobes by 50%, and I return an array factor that does that, then you can then turn around and hand me that array factor as input and demand that I reduce the side-lobes 50% relative to _that_; iterate several more times to reduce the side-lobes arbitrarily low, iterate to the point of arithmetic underflow to get a configuration that computationally cannot be differentiated from having eliminated the side-lobes. If you have an existence proof that reducing the side-lobes by 50% can always be done for any given configuration, then by induction the side-lobes can be reduced arbitrarily much, to below the uncertainty in your measurements. Ergo, one might as well just have a routine that, given the array configuration, reduces the side lobes to 0 (or below a given tolerance) directly.

It does not sound likely to me that this kind of complete side-lobe removal for an arbitrary number and placement and orientation of antenna can be achieved. Conversely, if one could establish that it could not be achieved for some given configuration, then that would be proof that it was not generally possible to reduce by 50% in all cases.

[Gupteswar]

Thank you robertson, for your valuable information. i need side lobe supression only. it does not need to be 50%. you asked me about % so i told you 50%. can you help in developing a code for this purpose?

[Walter Roberson]

If you had said that the side-lobes had to be suppressed by 0.1%, then the response would have been the same, that your request would be the equivalent of requiring that the side lobes be completely eliminated. 0.999^N might not approach 0 as quickly as 0.5^N does but it gets to the same suppression values eventually.

You question has the premise that given an arbitrary antenna configuration, that the side-lobes can be suppressed by some non-zero amount. But the configuration with the side lobes suppressed by the algorithm would be a new arbitrary antenna configuration, and if that had been the original configuration you would expect that it would still be able to have the side lobes suppressed by that non-zero amount. Just keep inputting the output configurations in to the program to achieve any level of suppression, if your premise is true.

The opposite of the premise is that there may be input configurations that already have their side-lobes suppressed down to the point that the "array factor" technique is not able to increase the suppression. There must be a possibility of failure.