Deconvolution of accumulated values
$begingroup$
Introduction
I have an unknown function of time $f(t)$ that I would like to learn based on experimental observations. I have an observable $g(t)$, which, to the best of my knowledge, is given by the convolution of $f(t)$
$g(t) = int_{0}^{infty}f(t-tau)h(tau)dtau + nu(t)$
where $h(t)$ is the known filter function, and $nu(t)$ is assumed to be Gaussian noise of known variance. $h(t)$ is typically an exponentially decaying function or similar. Let us assume for simplicity that we have measured $g(t)$ in the time interval $[0,T]$ and that $f(t)=g(t)=0$ for all $t<0$
Now, my understanding is that this problem is solved in practice by discretizing it, assuming $f(t)$ to be piecewise-constant over small time intervals $Delta t$. If we define
$f(t) = f_i$ where $i = lfloor frac{t}{Delta t} rfloor$
- $g_i = g(iDelta t)$
- $nu_i = nu(iDelta t)$
- $h_i = int_{i Delta t}^{(i+1)Delta t}h(t)dt$
the convolution integral can be rewritten in its discrete form as
$g_i = sum_{j=0}^{i} f_{i-j} h_j + nu_i$
It is then possible to solve the discrete problem by matrix inversion, fourier transform, etc...
Actual question
A central point in the above approximation is that the exact value of $g(t)$ is known at discrete intervals. In practice, however, the value of $g$ is accumulated over small time intervals, as opposed to being measured instantly. Namely, the actual discrete observable one obtains from the device is
$tilde{g}_i = frac{1}{Delta t} int_{i Delta t}^{(i+1)Delta t} g(t)dt$
I have tried to repeat the procedure to rewrite $tilde{g}_i$ as a function of $f_i$. However, it appears to be a page-long tedious calculation, and I keep making mistakes. Surely, I am not the first one to come across this problem. If you can provide a link to the derivation of the above equation, or perhaps its name or keyword if such exists, I would be grateful
discrete-optimization deconvolution
asked Jan 9 at 14:42
Aleksejs FominsAleksejs Fomins
480211
$endgroup$
add a comment |
$begingroup$
Introduction
I have an unknown function of time $f(t)$ that I would like to learn based on experimental observations. I have an observable $g(t)$, which, to the best of my knowledge, is given by the convolution of $f(t)$
$g(t) = int_{0}^{infty}f(t-tau)h(tau)dtau + nu(t)$
where $h(t)$ is the known filter function, and $nu(t)$ is assumed to be Gaussian noise of known variance. $h(t)$ is typically an exponentially decaying function or similar. Let us assume for simplicity that we have measured $g(t)$ in the time interval $[0,T]$ and that $f(t)=g(t)=0$ for all $t<0$
Now, my understanding is that this problem is solved in practice by discretizing it, assuming $f(t)$ to be piecewise-constant over small time intervals $Delta t$. If we define
$f(t) = f_i$ where $i = lfloor frac{t}{Delta t} rfloor$
- $g_i = g(iDelta t)$
- $nu_i = nu(iDelta t)$
- $h_i = int_{i Delta t}^{(i+1)Delta t}h(t)dt$
the convolution integral can be rewritten in its discrete form as
$g_i = sum_{j=0}^{i} f_{i-j} h_j + nu_i$
It is then possible to solve the discrete problem by matrix inversion, fourier transform, etc...
Actual question
A central point in the above approximation is that the exact value of $g(t)$ is known at discrete intervals. In practice, however, the value of $g$ is accumulated over small time intervals, as opposed to being measured instantly. Namely, the actual discrete observable one obtains from the device is
$tilde{g}_i = frac{1}{Delta t} int_{i Delta t}^{(i+1)Delta t} g(t)dt$
I have tried to repeat the procedure to rewrite $tilde{g}_i$ as a function of $f_i$. However, it appears to be a page-long tedious calculation, and I keep making mistakes. Surely, I am not the first one to come across this problem. If you can provide a link to the derivation of the above equation, or perhaps its name or keyword if such exists, I would be grateful
discrete-optimization deconvolution
asked Jan 9 at 14:42
Aleksejs FominsAleksejs Fomins
480211
$endgroup$
add a comment |
$begingroup$
Introduction
I have an unknown function of time $f(t)$ that I would like to learn based on experimental observations. I have an observable $g(t)$, which, to the best of my knowledge, is given by the convolution of $f(t)$
$g(t) = int_{0}^{infty}f(t-tau)h(tau)dtau + nu(t)$
where $h(t)$ is the known filter function, and $nu(t)$ is assumed to be Gaussian noise of known variance. $h(t)$ is typically an exponentially decaying function or similar. Let us assume for simplicity that we have measured $g(t)$ in the time interval $[0,T]$ and that $f(t)=g(t)=0$ for all $t<0$
Now, my understanding is that this problem is solved in practice by discretizing it, assuming $f(t)$ to be piecewise-constant over small time intervals $Delta t$. If we define
$f(t) = f_i$ where $i = lfloor frac{t}{Delta t} rfloor$
- $g_i = g(iDelta t)$
- $nu_i = nu(iDelta t)$
- $h_i = int_{i Delta t}^{(i+1)Delta t}h(t)dt$
the convolution integral can be rewritten in its discrete form as
$g_i = sum_{j=0}^{i} f_{i-j} h_j + nu_i$
It is then possible to solve the discrete problem by matrix inversion, fourier transform, etc...
Actual question
A central point in the above approximation is that the exact value of $g(t)$ is known at discrete intervals. In practice, however, the value of $g$ is accumulated over small time intervals, as opposed to being measured instantly. Namely, the actual discrete observable one obtains from the device is
$tilde{g}_i = frac{1}{Delta t} int_{i Delta t}^{(i+1)Delta t} g(t)dt$
I have tried to repeat the procedure to rewrite $tilde{g}_i$ as a function of $f_i$. However, it appears to be a page-long tedious calculation, and I keep making mistakes. Surely, I am not the first one to come across this problem. If you can provide a link to the derivation of the above equation, or perhaps its name or keyword if such exists, I would be grateful
discrete-optimization deconvolution
asked Jan 9 at 14:42
Aleksejs FominsAleksejs Fomins
480211
$endgroup$
Introduction
I have an unknown function of time $f(t)$ that I would like to learn based on experimental observations. I have an observable $g(t)$, which, to the best of my knowledge, is given by the convolution of $f(t)$
$g(t) = int_{0}^{infty}f(t-tau)h(tau)dtau + nu(t)$
where $h(t)$ is the known filter function, and $nu(t)$ is assumed to be Gaussian noise of known variance. $h(t)$ is typically an exponentially decaying function or similar. Let us assume for simplicity that we have measured $g(t)$ in the time interval $[0,T]$ and that $f(t)=g(t)=0$ for all $t<0$
Now, my understanding is that this problem is solved in practice by discretizing it, assuming $f(t)$ to be piecewise-constant over small time intervals $Delta t$. If we define
$f(t) = f_i$ where $i = lfloor frac{t}{Delta t} rfloor$
- $g_i = g(iDelta t)$
- $nu_i = nu(iDelta t)$
- $h_i = int_{i Delta t}^{(i+1)Delta t}h(t)dt$
the convolution integral can be rewritten in its discrete form as
$g_i = sum_{j=0}^{i} f_{i-j} h_j + nu_i$
It is then possible to solve the discrete problem by matrix inversion, fourier transform, etc...
Actual question
A central point in the above approximation is that the exact value of $g(t)$ is known at discrete intervals. In practice, however, the value of $g$ is accumulated over small time intervals, as opposed to being measured instantly. Namely, the actual discrete observable one obtains from the device is
$tilde{g}_i = frac{1}{Delta t} int_{i Delta t}^{(i+1)Delta t} g(t)dt$
I have tried to repeat the procedure to rewrite $tilde{g}_i$ as a function of $f_i$. However, it appears to be a page-long tedious calculation, and I keep making mistakes. Surely, I am not the first one to come across this problem. If you can provide a link to the derivation of the above equation, or perhaps its name or keyword if such exists, I would be grateful
discrete-optimization deconvolution
discrete-optimization deconvolution
asked Jan 9 at 14:42
Aleksejs FominsAleksejs Fomins
480211
asked Jan 9 at 14:42
Aleksejs FominsAleksejs Fomins
480211
asked Jan 9 at 14:42
Aleksejs FominsAleksejs Fomins
480211
asked Jan 9 at 14:42
Aleksejs FominsAleksejs Fomins
480211
asked Jan 9 at 14:42
Aleksejs FominsAleksejs Fomins
480211
480211
add a comment |
add a comment |
0
active
oldest
votes
Your Answer
StackExchange.ifUsing("editor", function () {
return StackExchange.using("mathjaxEditing", function () {
StackExchange.MarkdownEditor.creationCallbacks.add(function (editor, postfix) {
StackExchange.mathjaxEditing.prepareWmdForMathJax(editor, postfix, [["$", "$"], ["\\(","\\)"]]);
});
});
}, "mathjax-editing");
StackExchange.ready(function() {
var channelOptions = {
tags: "".split(" "),
id: "69"
};
initTagRenderer("".split(" "), "".split(" "), channelOptions);
StackExchange.using("externalEditor", function() {
// Have to fire editor after snippets, if snippets enabled
if (StackExchange.settings.snippets.snippetsEnabled) {
StackExchange.using("snippets", function() {
createEditor();
});
}
else {
createEditor();
}
});
function createEditor() {
StackExchange.prepareEditor({
heartbeatType: 'answer',
autoActivateHeartbeat: false,
convertImagesToLinks: true,
noModals: true,
showLowRepImageUploadWarning: true,
reputationToPostImages: 10,
bindNavPrevention: true,
postfix: "",
imageUploader: {
brandingHtml: "Powered by u003ca class="icon-imgur-white" href="https://imgur.com/"u003eu003c/au003e",
contentPolicyHtml: "User contributions licensed under u003ca href="https://creativecommons.org/licenses/by-sa/3.0/"u003ecc by-sa 3.0 with attribution requiredu003c/au003e u003ca href="https://stackoverflow.com/legal/content-policy"u003e(content policy)u003c/au003e",
allowUrls: true
},
noCode: true, onDemand: true,
discardSelector: ".discard-answer"
,immediatelyShowMarkdownHelp:true
});
}
});
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3067520%2fdeconvolution-of-accumulated-values%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
0
active
oldest
votes
0
active
oldest
votes
active
oldest
votes
active
oldest
votes
Thanks for contributing an answer to Mathematics Stack Exchange!
- Please be sure to answer the question. Provide details and share your research!
But avoid …
- Asking for help, clarification, or responding to other answers.
- Making statements based on opinion; back them up with references or personal experience.
Use MathJax to format equations. MathJax reference.
To learn more, see our tips on writing great answers.
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
StackExchange.ready(
function () {
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f3067520%2fdeconvolution-of-accumulated-values%23new-answer', 'question_page');
}
);
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Sign up or log in
StackExchange.ready(function () {
StackExchange.helpers.onClickDraftSave('#login-link');
});
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
Required, but never shown
