Where is the error in this naive derivation of the Kalman filter?
$begingroup$
The Kalman filter is a method of predicting the future state of a linear state space system based on the previous ones.
A linear, discrete-time, stationary, state-space model is a pair of real valued stochastic processes ${X_t }_{t in mathbb{N}},{Y_t}_{t in mathbb{N}}$ that obey the recursive equations
$$
begin{cases}
X_{t+1} = F X_t + v_t &t=0,1,2dots \
Y_t = H X_t + w_t &t=0,1,2dots
end{cases}
$$
where:
$F in mathbb{R}, H in mathbb{R}$
$v_t, w_t$ are random variables (additive noise) which admit a PDF (Probability Density Function);
The noise terms are zero-mean and:
$$
forall t_1 neq t_2 : v_{t_1} perp v_{t_2}, w_{t_1} perp w_{t_2} \
forall t_1, t_2 : v_{t_1} perp w_{t_2} \
forall t : E[v_t^2] = Q, E[w_t^2] = R
$$
where the simbol $X perp Y$ means that $X$ and $Y$ are independent and $Q,R$ are assumed to be positive real numbers. The initial condition of the recursion $x_0$ is a fixed real number.
Setting $Y^t = { Y_0, dots , Y_t }$The Kalman filter tells us that
$$E[X_t | Y^{t-1}] = alpha E[X_{t-1} | Y^{t-2}] + beta Y_{t-1} $$
where $alpha$ and $beta$ are carefully chosen and depend on the parameters of the linear model under consideration.
My question:
If I where to compute $E[X_t | Y^{t-1}]$ I would notice that $X_t = F X_{t-1} + v_{t-1} $ and $X_{t-1} = Y_{t-1}/ H - w_{t-1}/H$ so
$$X_t = F X_{t-1} + v_{t-1} = F (Y_{t-1}/ H - w_{t-1}/H) + v_{t-1} $$
and because $w_{t-1}, v_{t-1}$ are zero mean indipendent random variables
$$E[X_t | Y^{t-1}] = E[ F (Y_{t-1}/ H - w_{t-1}/H) + v_{t-1} | Y^{t-1}] = E[F Y_{t-1}/ H | Y^{t-1}] = F Y_{t-1}/ H $$
(I have also assumed $w_t$ is symmetric) then obviously by doing the same calculations $E[X_{t+1} | Y^{t}] = F Y_{t}/ H$.
There is no way to combine these two conditional expectations to obtain the recursive formula proven by Kalman.
Where is my mistake?
probability probability-theory statistics stochastic-processes time-series
asked Dec 22 '18 at 0:34
MonoliteMonolite
1,5552926
$endgroup$
add a comment |
$begingroup$
The Kalman filter is a method of predicting the future state of a linear state space system based on the previous ones.
A linear, discrete-time, stationary, state-space model is a pair of real valued stochastic processes ${X_t }_{t in mathbb{N}},{Y_t}_{t in mathbb{N}}$ that obey the recursive equations
$$
begin{cases}
X_{t+1} = F X_t + v_t &t=0,1,2dots \
Y_t = H X_t + w_t &t=0,1,2dots
end{cases}
$$
where:
$F in mathbb{R}, H in mathbb{R}$
$v_t, w_t$ are random variables (additive noise) which admit a PDF (Probability Density Function);
The noise terms are zero-mean and:
$$
forall t_1 neq t_2 : v_{t_1} perp v_{t_2}, w_{t_1} perp w_{t_2} \
forall t_1, t_2 : v_{t_1} perp w_{t_2} \
forall t : E[v_t^2] = Q, E[w_t^2] = R
$$
where the simbol $X perp Y$ means that $X$ and $Y$ are independent and $Q,R$ are assumed to be positive real numbers. The initial condition of the recursion $x_0$ is a fixed real number.
Setting $Y^t = { Y_0, dots , Y_t }$The Kalman filter tells us that
$$E[X_t | Y^{t-1}] = alpha E[X_{t-1} | Y^{t-2}] + beta Y_{t-1} $$
where $alpha$ and $beta$ are carefully chosen and depend on the parameters of the linear model under consideration.
My question:
If I where to compute $E[X_t | Y^{t-1}]$ I would notice that $X_t = F X_{t-1} + v_{t-1} $ and $X_{t-1} = Y_{t-1}/ H - w_{t-1}/H$ so
$$X_t = F X_{t-1} + v_{t-1} = F (Y_{t-1}/ H - w_{t-1}/H) + v_{t-1} $$
and because $w_{t-1}, v_{t-1}$ are zero mean indipendent random variables
$$E[X_t | Y^{t-1}] = E[ F (Y_{t-1}/ H - w_{t-1}/H) + v_{t-1} | Y^{t-1}] = E[F Y_{t-1}/ H | Y^{t-1}] = F Y_{t-1}/ H $$
(I have also assumed $w_t$ is symmetric) then obviously by doing the same calculations $E[X_{t+1} | Y^{t}] = F Y_{t}/ H$.
There is no way to combine these two conditional expectations to obtain the recursive formula proven by Kalman.
Where is my mistake?
probability probability-theory statistics stochastic-processes time-series
asked Dec 22 '18 at 0:34
MonoliteMonolite
1,5552926
$endgroup$
1
$begingroup$
$mathsf{E}[w_{t-1}mid Y^{t-1}]=0$?
$endgroup$
– d.k.o.
Dec 22 '18 at 6:59
$begingroup$
@d.k.o. thanks, of course!
$endgroup$
– Monolite
Dec 22 '18 at 11:11
add a comment |
$begingroup$
The Kalman filter is a method of predicting the future state of a linear state space system based on the previous ones.
A linear, discrete-time, stationary, state-space model is a pair of real valued stochastic processes ${X_t }_{t in mathbb{N}},{Y_t}_{t in mathbb{N}}$ that obey the recursive equations
$$
begin{cases}
X_{t+1} = F X_t + v_t &t=0,1,2dots \
Y_t = H X_t + w_t &t=0,1,2dots
end{cases}
$$
where:
$F in mathbb{R}, H in mathbb{R}$
$v_t, w_t$ are random variables (additive noise) which admit a PDF (Probability Density Function);
The noise terms are zero-mean and:
$$
forall t_1 neq t_2 : v_{t_1} perp v_{t_2}, w_{t_1} perp w_{t_2} \
forall t_1, t_2 : v_{t_1} perp w_{t_2} \
forall t : E[v_t^2] = Q, E[w_t^2] = R
$$
where the simbol $X perp Y$ means that $X$ and $Y$ are independent and $Q,R$ are assumed to be positive real numbers. The initial condition of the recursion $x_0$ is a fixed real number.
Setting $Y^t = { Y_0, dots , Y_t }$The Kalman filter tells us that
$$E[X_t | Y^{t-1}] = alpha E[X_{t-1} | Y^{t-2}] + beta Y_{t-1} $$
where $alpha$ and $beta$ are carefully chosen and depend on the parameters of the linear model under consideration.
My question:
If I where to compute $E[X_t | Y^{t-1}]$ I would notice that $X_t = F X_{t-1} + v_{t-1} $ and $X_{t-1} = Y_{t-1}/ H - w_{t-1}/H$ so
$$X_t = F X_{t-1} + v_{t-1} = F (Y_{t-1}/ H - w_{t-1}/H) + v_{t-1} $$
and because $w_{t-1}, v_{t-1}$ are zero mean indipendent random variables
$$E[X_t | Y^{t-1}] = E[ F (Y_{t-1}/ H - w_{t-1}/H) + v_{t-1} | Y^{t-1}] = E[F Y_{t-1}/ H | Y^{t-1}] = F Y_{t-1}/ H $$
(I have also assumed $w_t$ is symmetric) then obviously by doing the same calculations $E[X_{t+1} | Y^{t}] = F Y_{t}/ H$.
There is no way to combine these two conditional expectations to obtain the recursive formula proven by Kalman.
Where is my mistake?
probability probability-theory statistics stochastic-processes time-series
asked Dec 22 '18 at 0:34
MonoliteMonolite
1,5552926
$endgroup$
The Kalman filter is a method of predicting the future state of a linear state space system based on the previous ones.
A linear, discrete-time, stationary, state-space model is a pair of real valued stochastic processes ${X_t }_{t in mathbb{N}},{Y_t}_{t in mathbb{N}}$ that obey the recursive equations
$$
begin{cases}
X_{t+1} = F X_t + v_t &t=0,1,2dots \
Y_t = H X_t + w_t &t=0,1,2dots
end{cases}
$$
where:
$F in mathbb{R}, H in mathbb{R}$
$v_t, w_t$ are random variables (additive noise) which admit a PDF (Probability Density Function);
The noise terms are zero-mean and:
$$
forall t_1 neq t_2 : v_{t_1} perp v_{t_2}, w_{t_1} perp w_{t_2} \
forall t_1, t_2 : v_{t_1} perp w_{t_2} \
forall t : E[v_t^2] = Q, E[w_t^2] = R
$$
where the simbol $X perp Y$ means that $X$ and $Y$ are independent and $Q,R$ are assumed to be positive real numbers. The initial condition of the recursion $x_0$ is a fixed real number.
Setting $Y^t = { Y_0, dots , Y_t }$The Kalman filter tells us that
$$E[X_t | Y^{t-1}] = alpha E[X_{t-1} | Y^{t-2}] + beta Y_{t-1} $$
where $alpha$ and $beta$ are carefully chosen and depend on the parameters of the linear model under consideration.
My question:
If I where to compute $E[X_t | Y^{t-1}]$ I would notice that $X_t = F X_{t-1} + v_{t-1} $ and $X_{t-1} = Y_{t-1}/ H - w_{t-1}/H$ so
$$X_t = F X_{t-1} + v_{t-1} = F (Y_{t-1}/ H - w_{t-1}/H) + v_{t-1} $$
and because $w_{t-1}, v_{t-1}$ are zero mean indipendent random variables
$$E[X_t | Y^{t-1}] = E[ F (Y_{t-1}/ H - w_{t-1}/H) + v_{t-1} | Y^{t-1}] = E[F Y_{t-1}/ H | Y^{t-1}] = F Y_{t-1}/ H $$
(I have also assumed $w_t$ is symmetric) then obviously by doing the same calculations $E[X_{t+1} | Y^{t}] = F Y_{t}/ H$.
There is no way to combine these two conditional expectations to obtain the recursive formula proven by Kalman.
Where is my mistake?
probability probability-theory statistics stochastic-processes time-series
probability probability-theory statistics stochastic-processes time-series
asked Dec 22 '18 at 0:34
MonoliteMonolite
1,5552926
asked Dec 22 '18 at 0:34
MonoliteMonolite
1,5552926
asked Dec 22 '18 at 0:34
MonoliteMonolite
1,5552926
asked Dec 22 '18 at 0:34
MonoliteMonolite
1,5552926
asked Dec 22 '18 at 0:34
MonoliteMonolite
1,5552926
1,5552926
1
$begingroup$
$mathsf{E}[w_{t-1}mid Y^{t-1}]=0$?
$endgroup$
– d.k.o.
Dec 22 '18 at 6:59
$begingroup$
@d.k.o. thanks, of course!
$endgroup$
– Monolite
Dec 22 '18 at 11:11
add a comment |
1
$begingroup$
$mathsf{E}[w_{t-1}mid Y^{t-1}]=0$?
$endgroup$
– d.k.o.
Dec 22 '18 at 6:59
$begingroup$
@d.k.o. thanks, of course!
$endgroup$
– Monolite
Dec 22 '18 at 11:11
1
1
$begingroup$
$mathsf{E}[w_{t-1}mid Y^{t-1}]=0$?
$endgroup$
– d.k.o.
Dec 22 '18 at 6:59
$begingroup$
$mathsf{E}[w_{t-1}mid Y^{t-1}]=0$?
$endgroup$
– d.k.o.
Dec 22 '18 at 6:59
$begingroup$
@d.k.o. thanks, of course!
$endgroup$
– Monolite
Dec 22 '18 at 11:11
$begingroup$
@d.k.o. thanks, of course!
$endgroup$
– Monolite
Dec 22 '18 at 11:11
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%2f3049037%2fwhere-is-the-error-in-this-naive-derivation-of-the-kalman-filter%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%2f3049037%2fwhere-is-the-error-in-this-naive-derivation-of-the-kalman-filter%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
1
$begingroup$
$mathsf{E}[w_{t-1}mid Y^{t-1}]=0$?
$endgroup$
– d.k.o.
Dec 22 '18 at 6:59
$begingroup$
@d.k.o. thanks, of course!
$endgroup$
– Monolite
Dec 22 '18 at 11:11