Forums

help in filter coefficients

Started by asik nisha January 30, 2012
hi

      i'm doing my project in genetic algorithm.... i need to find the
best filter coefficient value using this genetic algorithm..... i
already designed one adaptive filter and also find its
coefficients...... So by using GA the mean square value of the filter
coefficient is calculate and compare the error values and find the
best value among it.... i tried it in optimization tool... i cant get
a perfect result... can anyone have any ideas regarding this....
thanks in advance....
On 1/30/2012 11:28 AM, asik nisha wrote:
> hi > > i'm doing my project in genetic algorithm.... i need to find the > best filter coefficient value using this genetic algorithm..... i > already designed one adaptive filter and also find its > coefficients...... So by using GA the mean square value of the filter > coefficient is calculate and compare the error values and find the > best value among it.... i tried it in optimization tool... i cant get > a perfect result... can anyone have any ideas regarding this.... > thanks in advance....
It seems to me that, while filter design may be a good exercise for using a genetic algorithm, genetic algorithms are not especially good for designing filters. For filter design, ask here. For programming paradigms, look elsewhere. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
On Mon, 30 Jan 2012 08:28:13 -0800, asik nisha wrote:

> hi > > i'm doing my project in genetic algorithm.... i need to find the > best filter coefficient value using this genetic algorithm..... i > already designed one adaptive filter and also find its > coefficients...... So by using GA the mean square value of the filter > coefficient is calculate and compare the error values and find the best > value among it.... i tried it in optimization tool... i cant get a > perfect result... can anyone have any ideas regarding this.... thanks in > advance....
Much is not clear... Are you saying that you are trying to find the coefficients for an optimal fixed-coefficient filter, and are attempting to do so first by building an adaptive filter and letting it run, then by attempting to find the coefficients through a genetic algorithm? Or are you trying to find a good adaptive filter, and are attempting to find the best twiddle factors for it by using a genetic algorithm? I deduce that you are defining optimum as the minimum mean square error. You complain that you can't find the 'perfect' filter. What do you mean by 'perfect'? Do you have reason to believe that this 'perfect' filter exists? It is a rare case in the natural world that one can filter signal +noise and get back a perfect version of the signal -- it is much more common that one's filter gives back an approximation, which is why signal processing types work at building "optimal", rather than "perfect" filters. -- Tim Wescott Control system and signal processing consulting www.wescottdesign.com
On 30.01.2012 17:28, asik nisha wrote:
> hi > > i'm doing my project in genetic algorithm.... i need to find the > best filter coefficient value using this genetic algorithm..... i > already designed one adaptive filter and also find its > coefficients...... So by using GA the mean square value of the filter > coefficient is calculate and compare the error values and find the > best value among it.... i tried it in optimization tool... i cant get > a perfect result... can anyone have any ideas regarding this.... > thanks in advance....
What kind of filter? For FIR, a very simple approach worked quite fine for me... Andre
On Jan 31, 1:40&#2013266080;pm, Andre <lod...@pathme.de> wrote:
> On 30.01.2012 17:28, asik nisha wrote: > > > hi > > > &#2013266080; &#2013266080; &#2013266080; &#2013266080;i'm doing my project in genetic algorithm.... i need to find the > > best filter coefficient value using this genetic algorithm..... i > > already designed one adaptive filter and also find its > > coefficients...... So by using GA the mean square value of the filter > > coefficient is calculate and compare the error values and find the > > best value among it.... i tried it in optimization tool... i cant get > > a perfect result... can anyone have any ideas regarding this.... > > thanks in advance.... > > What kind of filter? > > For FIR, a very simple approach worked quite fine for me... > > Andre
its a adaptive FIR filter....
On Tue, 31 Jan 2012 05:40:28 -0800, ABU BENAZIR wrote:

> On Jan 31, 1:40&nbsp;pm, Andre <lod...@pathme.de> wrote: >> On 30.01.2012 17:28, asik nisha wrote: >> >> > hi >> >> > &nbsp; &nbsp; &nbsp; &nbsp;i'm doing my project in genetic algorithm.... i need to >> > &nbsp; &nbsp; &nbsp; &nbsp;find the >> > best filter coefficient value using this genetic algorithm..... i >> > already designed one adaptive filter and also find its >> > coefficients...... So by using GA the mean square value of the filter >> > coefficient is calculate and compare the error values and find the >> > best value among it.... i tried it in optimization tool... i cant get >> > a perfect result... can anyone have any ideas regarding this.... >> > thanks in advance.... >> >> What kind of filter? >> >> For FIR, a very simple approach worked quite fine for me... >> >> Andre > > its a adaptive FIR filter....
It's still not clear what parameters you're trying to change with the genetic algorithm. If you want information, you have to give enough information in your question to get an answer. Are you trying to use the GA to change the adaptation parameters? Are you trying to use the GA to find the actual filter parameters? What? Give an _outline_ of what you're doing. -- Tim Wescott Control system and signal processing consulting www.wescottdesign.com
On 1/31/2012 11:08 AM, Tim Wescott wrote:
> On Tue, 31 Jan 2012 05:40:28 -0800, ABU BENAZIR wrote: > >> On Jan 31, 1:40 pm, Andre<lod...@pathme.de> wrote: >>> On 30.01.2012 17:28, asik nisha wrote: >>> >>>> hi >>> >>>> i'm doing my project in genetic algorithm.... i need to >>>> find the >>>> best filter coefficient value using this genetic algorithm..... i >>>> already designed one adaptive filter and also find its >>>> coefficients...... So by using GA the mean square value of the filter >>>> coefficient is calculate and compare the error values and find the >>>> best value among it.... i tried it in optimization tool... i cant get >>>> a perfect result... can anyone have any ideas regarding this.... >>>> thanks in advance.... >>> >>> What kind of filter? >>> >>> For FIR, a very simple approach worked quite fine for me... >>> >>> Andre >> >> its a adaptive FIR filter.... > > It's still not clear what parameters you're trying to change with the > genetic algorithm. If you want information, you have to give enough > information in your question to get an answer. > > Are you trying to use the GA to change the adaptation parameters? Are > you trying to use the GA to find the actual filter parameters? What? > Give an _outline_ of what you're doing.
Also, state the purpose. Are you trying to make a GA work in order to learn about GAs? Are you trying to design a filter, believing that a GA is the best approach? Some other intention? Jerry -- Engineering is the art of making what you want from things you can get. &macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;&macr;
On Tue, 31 Jan 2012 12:41:22 -0500, Jerry Avins wrote:

> On 1/31/2012 11:08 AM, Tim Wescott wrote: >> On Tue, 31 Jan 2012 05:40:28 -0800, ABU BENAZIR wrote: >> >>> On Jan 31, 1:40 pm, Andre<lod...@pathme.de> wrote: >>>> On 30.01.2012 17:28, asik nisha wrote: >>>> >>>>> hi >>>> >>>>> i'm doing my project in genetic algorithm.... i need to find >>>>> the >>>>> best filter coefficient value using this genetic algorithm..... i >>>>> already designed one adaptive filter and also find its >>>>> coefficients...... So by using GA the mean square value of the >>>>> filter coefficient is calculate and compare the error values and >>>>> find the best value among it.... i tried it in optimization tool... >>>>> i cant get a perfect result... can anyone have any ideas regarding >>>>> this.... thanks in advance.... >>>> >>>> What kind of filter? >>>> >>>> For FIR, a very simple approach worked quite fine for me... >>>> >>>> Andre >>> >>> its a adaptive FIR filter.... >> >> It's still not clear what parameters you're trying to change with the >> genetic algorithm. If you want information, you have to give enough >> information in your question to get an answer. >> >> Are you trying to use the GA to change the adaptation parameters? Are >> you trying to use the GA to find the actual filter parameters? What? >> Give an _outline_ of what you're doing. > > Also, state the purpose. Are you trying to make a GA work in order to > learn about GAs? Are you trying to design a filter, believing that a GA > is the best approach? Some other intention?
I get the impression he's a student -- he may be doing it because his prof decided to mess with GA-designed filters, but wanted a student to actually do the work. -- My liberal friends think I'm a conservative kook. My conservative friends think I'm a liberal kook. Why am I not happy that they have found common ground? Tim Wescott, Communications, Control, Circuits & Software http://www.wescottdesign.com
Tim Wescott <tim@seemywebsite.com> wrote:
> On Tue, 31 Jan 2012 12:41:22 -0500, Jerry Avins wrote:
(snip)
>> Also, state the purpose. Are you trying to make a GA work in order to >> learn about GAs? Are you trying to design a filter, believing that a GA >> is the best approach? Some other intention?
> I get the impression he's a student -- he may be doing it because his > prof decided to mess with GA-designed filters, but wanted a student to > actually do the work.
The first time I heard about Genetic Algorithms was in a talk reporting that they didn't work well for analysis of genome data. That was early in the years of the human genome project, when finding new ways to analyze large sets of sequence data was important. It seems likely that they aren't very good for filter design, either, but someone can still try. -- glen
On Jan 31, 3:47=A0pm, glen herrmannsfeldt <g...@ugcs.caltech.edu> wrote:

> ... > The first time I heard about Genetic Algorithms was in a talk reporting > that they didn't work well for analysis of genome data. That was early > in the years of the human genome project, when finding new ways to analyz=
e
> large sets of sequence data was important. It seems likely that they aren=
't
> very good for filter design, either, but someone can still try. > ...
Genetic algorithms have been used successfully in the design of filters for decades. Perhaps you would learn more about the capabilities of tools by looking for those who have used them successfully than you have from an anecdote about those who have failed to use the tools successfully. Should you care to know about the topic about which you speak try Google: filter design by genetic algorithm Dale B. Dalrymple