当前位置:天才代写 > 数学代写代考,北美/加拿大/英国靠谱的数学作业代写机构 > 滤波器论文代写 数学作业代写

滤波器论文代写 数学作业代写

2021-09-12 17:10 星期日 所属: 数学代写代考,北美/加拿大/英国靠谱的数学作业代写机构 浏览:460

滤波器论文代写

Research-Linked Topics: Continuous-time Linear Filter

滤波器论文代写 Your overall aim is to write a good essay about the filtering problem. You will find some possible tasks, in the form of exercises in Section 2.

1 What to do

Your overall aim is to write a good essay about the filtering problem. You will find some possible tasks, in the form of exercises in Section 2. You are not expected to complete them all. Instead they should serve as something you can focus on in your essay. 滤波器论文代写

When writing your essay do not refer to the specific exercises, your aim should be to produce something that can be read on its own.  In other words do not treat this  as answering an exam where you write just the solutions.

1.1 Initial 2 page part

It is recommended that you  write  solutions  to  some  parts  of  Exercises  2.1,  2.2  and 2.3.

1.2 Main project

You should take into account the feedback received for the initial 2 page part. If the same mistakes appear uncorrected then you will certainly lose marks. It is recom- mended that you provide solutions to some of the remaining exercises. If you do less theoretical work then it is expected that you do more computational work (and that you provide careful analysis of the computational work). 滤波器论文代写

1.3 Reading

Regarding the recommended books: Bain and Crisan [1] is the most general and hence the hardest to read. Lipster and Shiryayev [4] contains many details but it may also be difficult to read. Fristedt, Jain and Krylov [2] is written very nicely at an introductory level (but may not be available in the library). Øksendal [3] is widely available and treats the linear case but differently to what was discussed in the lecture and it is not recommended.

滤波器论文代写
滤波器论文代写

2 Kalman–Bucy ftlter and exercises 滤波器论文代写

The continuous time linear filter is also known as the Kalman–Bucy filter. We fix a probability space (Ω, F, P), a time horizon T > 0.

The hidden “state” and “observation” processes are given as a stochastic differen- tial equation (SDE) system:

(state) dxt Fxtdt + σdWt ,x0 ∼ N (0, r0),(observation) dyt = Hxtdt + dVt ,Y0 = 0.

Here F, H ∈ R are given constants and x0 is assumed independent of (W, V ) while (W, V ) is an R2-valued Wiener process (which in turn means that W and V are independent). 滤波器论文代写

Aim:    Given a function ϕ : Rd2 → R find the best mean square estimate ϕˆ for ϕ(xt), given the observations up to time t i.e. given the path y[0,t] := (ys)s∈[0,t]. Hence we are  saying  that  ϕˆ must  satisfy:

E|ϕˆ − ϕ(xt)|2 = min E|f (y[0,t]) − ϕ(xt)|2 .

Exercise  2.1.

Show  that  ϕˆ = E[ϕ(xt)|σ(y[0,t])].

We now know that a conditional expectation gives the best mean square estimate but we do not know how to compute this. Let us write Yt := σ(y[0,t]) and let us define

Pt(ϕ) := E[ϕ(xt)|Yt] ∀ϕ bounded . 滤波器论文代写

Our aim is thus to be able to compute Pt for a given ϕ.  Note that we may view Pt as the law of xt conditional on Yt.

Exercise 2.2.

Consider a random variable X.

i)Define the law (or distribution) ofX.

ii)Show that  if  we  know  E[ϕ(X)]  for  any  ϕ : R → R bounded  and  measurable  then we know the law of X. 滤波器论文代写

iii)Showthat if we know the law of X then we can evaluate E[ϕ(X)] for any ϕ : R →

R bounded and measurable. Hint: Consider simple functions ϕ first.

Exercise 2.3.

Assume  that  the  observation  process  is  replaced  by  dyt =  dVt.  That is, the observations do not tell us anything about the process. Assume that ϕ ∈ C2.

  1. Show that  Pt(ϕ) = E[ϕ(xt)].
  2. Use Itˆo  formula  and  then  take  expectation  to  show  that

dPt(ϕ) = Pt(Lϕ)dt,  滤波器论文代写

where  L(x)ϕ(x) :=  1σ2ϕ(x) + Fxϕ(x).

3.Comment on this

Proposition 2.4 (Kalman–Bucy Filter). For the linear filtering problem here we have that Pt, which is the distribution of xt conditional on Yt is normal with conditional mean xˆt and conditional variance Rt, where

dxˆt F xˆtdt HRtd(yt − Hxˆtdt,

and Rt satisfies the ordinary dierential equation (ODE) 滤波器论文代写

xˆ0 =0

dRt = σ2 + 2FR dt t

H Rt , R0

= r0 .

Note that ODEs of this form are called Riccati equations.

Exercise 2.5.

Prove the first statement of Proposition 2.4. That is, show that Pt, which is the distribution of xt conditional on Yt is normal. Hint. There is a Lemma in Bain and Crisan [1] that makes a good starting point but you should include more detail e.g. by following the reference given there. 滤波器论文代写

Exercise 2.6.

Show that we have the equations for xˆt and Rt as claimed in Proposi- tion 2.4. You can do this by considering

(state) dxt b(xt, yt)dt + θ(xt, yt)dwt + ρ(xt, yt)dvt , x0 = ξ,

(observation) dyt B(xt, yt)dt dvt ,

Y0 = η .

Here xt ∈ Rd1 , yt ∈ Rd2 , (w, v) is a Rd′+d2 -Wiener process (in particular all components are independnt and so w and v are independent Wiener processes taking values in Rd and Rd′ respectively).  The initial points (ξ, η) are assumed independent of each other and (w, v). The functions  滤波器论文代写

b : Rd1+d2 → Rd, B : Rd1+d2 → Rd2 ,

θ : Rd1+d2 → Rd1×d, ρ : Rd1+d2 → Rd1×d2,

are assumed to be Lipschitz continuous.

Now follow these steps:

i)Let

to define a change of measure P¯  and use Girsanov’s theorem to show that yt − y0

is a Wiener process (assume B is bounded).

ii)Use Lemma  3.1  to  show  that  with  µt(ϕ) := E¯[ϕ(xtt−1|Yt]  we  have

P (ϕ) =  µt(ϕ) .  滤波器论文代写

t µt(1)

iii)Derive the Zakaiequation:

滤波器论文代写
滤波器论文代写

Hint: You can use Lemma 3.3 without proof.

iv)Derive the Kushner–Shiryayevequation:

dPt(ϕ) = Pt(Ltϕ)dt + (Pt(Mtϕ) − Pt(ϕ)Pt(Bt)) dV¯t , P0(ϕ) = E[ϕ(ξ)],

∀ϕ ∈ C2 . (1)  滤波器论文代写

Here  dV¯t := dyt − Pt(Bt)dt.  This  is  called  the  “innovation  process”.

Hint: First show that t(1) = µt(1)Pt(B)dyt by using Lemma 3.3 again.

v)UseExercise 5 and apply (1) with ϕ(x) = x and ϕ(x) = x2 to obtain the result of Proposition 2.4.

vi)Explainwhy using ϕ(x) = x (or x2) is not mathematically valid and suggest how you could overcome this.

Exercise 2.7.

Use a programming  language  of  your  choice  (though  Matlab,  Python or R would probably be the most suitable) to implement the Kalman–Bucy filter. You will need to carry out the following steps: fix T > 0, and N  ∈ N for the number of  time steps you will use in discretization. Let τ = T/N .

i)Simulate the stateas

Xi+1 = Xi FXiτ + σ√τ Zi , i = 0, . . . , N − 1.

Here Zi are independent samples from N (0, 1) (and independent from X0).

ii)Simulate the signalas  滤波器论文代写

Y i+1 = Y i HXiτ + √τ Z˜i , i = 0, . . . , N − 1.

Here  Z˜i  are  independent  samples  from  N (0, 1)  (and  independent  from  X0 and from Zi).

iii)0Solve the Riccati ODE

iv)Calculate xˆ  from  the  above.

You should be able to produce simulation results similar to Figure 1.

3 Additional Material 滤波器论文代写

We will need a Lemma on conditional expectations under a change of measure:

Lemma  3.1.  Take two probability measures and P¯  such that1 P¯  << with

dP¯  = ΛdP.  滤波器论文代写

Let  be  a  subσalgebra  of  F.  Then  P¯  almost  surely  E[Λ|G]  > 0.  Moreover  for  any

F-random variable X we have

滤波器论文代写
滤波器论文代写

where E¯  denotes expectation taken under .


1Recall  that  we  say  that  a  measure  P¯  is  absolutely  continuous  with  respect  to  a  measure  P  if P(E) = 0 implies that P¯(E) = 0.  We write P¯ << P.

Figure 1: Example of one run of the Kalman–Bucy filter from solution to Exercise 2.7. Note that the parameters taken were σ = 0.5, F  = 1, r0 = 105, H = 3/2, T  = 1,      N = 400.

Exercise 3.2.

Prove Lemma 3.1.

We also require the following Lemma, which we present without proof.

Lemma  3.3.  Fix  a  probability  space  (Ω, F, P¯).  Let  (W, Y )  be  a  d + d2dimensional Wiener  process.   Let  ξ, η be  random  variables  independent  of  (W, Y )  and  let  Yt := σ(η, Ys s ≤ t).  Let (ft)tT be a bounded stochastic process.  Then  

References 滤波器论文代写

[1]Bain and D. Crisan. Fundamentals of Stochastic Filtering. Springer, 2009.

[2]Fristedt and N. Jain and N. V. Krylov. Filtering and Prediction: A Primer. AMS,2007.

[3]Øksendal. Stochastic Dierential Equations. Springer 2003.

[4]S. Lipster and A. N. Shiryayev. Statistics of Random Processes. Springer, 2001.

滤波器论文代写
滤波器论文代写

其他代写:代写CS C++代写 java代写 matlab代写 web代写 app代写 作业代写 物理代写 数学代写 考试助攻 paper代写 金融经济统计代写 python代写 Exercise代写

合作平台:essay代写 论文代写 写手招聘 英国留学生代写

 

天才代写-代写联系方式