- #1
lep11
- 380
- 7
Homework Statement
I have a few data-analysis problems due to Thursday,
1. Assume that a sandbag is dropped at different heights and the observations are (z
i;ti) pairs. Physical model for a free fall is z=½gt2. Assume that the height measurement z has an additive random error v. The observation model is then z=h(θ;t) +v,where θ is the unknown parameter vector. Write the observation models (withpaper and pen) for the following cases by treating systematic errors as model parameters:
a) height z has a systematic error z0 (zobs=zreal+z0)
b) falling time thas a systematic error t0 (tobs=treal+t0)
c) both z and t have additive systematic errors. Simulate results by using Matlab for graphic presentation.
2. Assume that a sandbag is dropped at different heights and the observations are (zi,ti)pairs. It’s known that the clock advances p percents, but the clock can be considered precise (no random error for t). In addition, assume that the height z has a systematic error z0. Write the (linear) observation model withpaper and pen. How would you solve a value for g by using the model?
4. Simulate with
Matlab the distributions of:
y1=sin(x)
y2=cos(x)
y3=log(x)
when x is uniformly distributed between [−π, π] in Eq. (1) and in Eq. (2), and uniformly distributed
between [1,2] in Eq. (3).
5. The second order polynomial function is of the form
y=p(1)x2+(2)x+p(3). The exact roots of the polynomial are x1=1 and x2=2. For some reason, there is a uniformly distributed error in the observed polynomial coefficients p(1), p2) and p(3). Here, we examine the distribution of observed roots in that case. (Matlab simulation)
a) Solve the exact polynomial coefficients p(1), p(2) and p(3) (command poly) and plot the polynomial y
in range x∈[−1,...,4] (command polyval)
b) Generate N= 5000 observations for polynomial coefficients with uniformly distributed noise.
Solve the roots for each observation (command roots).
c) Plot a histogram of observed roots x1 and x2
The attempt at a solution
1.
a.) z=h(θ;t) +v +z0
b.) z=h(θ;t+t0) +v
c.) z=h(θ;t+t0) +v+z0
%% Problem 1/a
g=10; % gravity constant
h_err=2; % systematic error of height
t=[0.3 0.5 0.7 1 2 2.5 3 3.7 3.8 4]';
tt=[0:.1:4]';
h=0.5*g*t.^2 + h_err + 10*randn(size(t));
H=[0.5*t.^2 ones(size(t))];
th=H\h
HH=[0.5*tt.^2 ones(size(tt))];
figure(1),clf
plot(t,h,'+',tt,HH*th)
xlabel 't'
ylabel 'h'
%% Problem 1/b
g=10; % gravity constant
t_err=2; % systematic error of time
t=[0.3 0.5 0.7 1 2 2.5 3 3.7 3.8 4]';
tt=[0:.1:4]';
h=0.5*g*(t+t_err).^2 + 10*randn(size(t));
H=[0.5*t.^2 ones(size(t))];
th=H\h;
HH=[0.5*tt.^2 ones(size(tt))];
figure(2),clf
plot(t,h,'+',tt,HH*th)
xlabel 't'
ylabel 'h'
%% Problem 1/c
g=10; % gravity constant
h_err=2; % systematic error of height
t_err=0;5; % systematic error of time
t=[0.3 0.5 0.7 1 2 2.5 3 3.7 3.8 4]';
tt=[0:.1:4]';
h=0.5*g*(t+t_err).^2 + h_err + 10*randn(size(t));
H=[0.5*t.^2 ones(size(t))];
th=H\h;
HH=[0.5*tt.^2 ones(size(tt))];
figure(3),clf
plot(t,h,'+',tt,HH*th)
xlabel 't'
ylabel 'h'
%% Problem 2
g=10; % gravity constant
t_err=0;01*t; % systematic error of time, e.g.clock is advancing one percent
t=[0.3 0.5 0.7 1 2 2.5 3 3.7 3.8 4]';
tt=[0:.1:4]';
h=0.5*g*(t+t_err).^2 + h_err;
H=[0.5*t.^2 ones(size(t))];
th=H\h
HH=[0.5*tt.^2 ones(size(tt))];
figure(4),clf
plot(t,h,'+',tt,HH*th)
xlabel 't'
ylabel 'h'
%% Problem 4
x_1=unifdist(-pi,pi,2,1);
y_1=sin (x_1);
y_2=cos (x_1);x_3=unifdist(1,2,2,1);
y_3=log (x_3);
I will appreciate any help as I am not very familiar with MATLAB yet \\ could sb please check what I've done so far
p.s. not sure if this is the right forum section for matlab-based questions
I have a few data-analysis problems due to Thursday,
1. Assume that a sandbag is dropped at different heights and the observations are (z
i;ti) pairs. Physical model for a free fall is z=½gt2. Assume that the height measurement z has an additive random error v. The observation model is then z=h(θ;t) +v,where θ is the unknown parameter vector. Write the observation models (withpaper and pen) for the following cases by treating systematic errors as model parameters:
a) height z has a systematic error z0 (zobs=zreal+z0)
b) falling time thas a systematic error t0 (tobs=treal+t0)
c) both z and t have additive systematic errors. Simulate results by using Matlab for graphic presentation.
2. Assume that a sandbag is dropped at different heights and the observations are (zi,ti)pairs. It’s known that the clock advances p percents, but the clock can be considered precise (no random error for t). In addition, assume that the height z has a systematic error z0. Write the (linear) observation model withpaper and pen. How would you solve a value for g by using the model?
4. Simulate with
Matlab the distributions of:
y1=sin(x)
y2=cos(x)
y3=log(x)
when x is uniformly distributed between [−π, π] in Eq. (1) and in Eq. (2), and uniformly distributed
between [1,2] in Eq. (3).
5. The second order polynomial function is of the form
y=p(1)x2+(2)x+p(3). The exact roots of the polynomial are x1=1 and x2=2. For some reason, there is a uniformly distributed error in the observed polynomial coefficients p(1), p2) and p(3). Here, we examine the distribution of observed roots in that case. (Matlab simulation)
a) Solve the exact polynomial coefficients p(1), p(2) and p(3) (command poly) and plot the polynomial y
in range x∈[−1,...,4] (command polyval)
b) Generate N= 5000 observations for polynomial coefficients with uniformly distributed noise.
Solve the roots for each observation (command roots).
c) Plot a histogram of observed roots x1 and x2
The attempt at a solution
1.
a.) z=h(θ;t) +v +z0
b.) z=h(θ;t+t0) +v
c.) z=h(θ;t+t0) +v+z0
%% Problem 1/a
g=10; % gravity constant
h_err=2; % systematic error of height
t=[0.3 0.5 0.7 1 2 2.5 3 3.7 3.8 4]';
tt=[0:.1:4]';
h=0.5*g*t.^2 + h_err + 10*randn(size(t));
H=[0.5*t.^2 ones(size(t))];
th=H\h
HH=[0.5*tt.^2 ones(size(tt))];
figure(1),clf
plot(t,h,'+',tt,HH*th)
xlabel 't'
ylabel 'h'
%% Problem 1/b
g=10; % gravity constant
t_err=2; % systematic error of time
t=[0.3 0.5 0.7 1 2 2.5 3 3.7 3.8 4]';
tt=[0:.1:4]';
h=0.5*g*(t+t_err).^2 + 10*randn(size(t));
H=[0.5*t.^2 ones(size(t))];
th=H\h;
HH=[0.5*tt.^2 ones(size(tt))];
figure(2),clf
plot(t,h,'+',tt,HH*th)
xlabel 't'
ylabel 'h'
%% Problem 1/c
g=10; % gravity constant
h_err=2; % systematic error of height
t_err=0;5; % systematic error of time
t=[0.3 0.5 0.7 1 2 2.5 3 3.7 3.8 4]';
tt=[0:.1:4]';
h=0.5*g*(t+t_err).^2 + h_err + 10*randn(size(t));
H=[0.5*t.^2 ones(size(t))];
th=H\h;
HH=[0.5*tt.^2 ones(size(tt))];
figure(3),clf
plot(t,h,'+',tt,HH*th)
xlabel 't'
ylabel 'h'
%% Problem 2
g=10; % gravity constant
t_err=0;01*t; % systematic error of time, e.g.clock is advancing one percent
t=[0.3 0.5 0.7 1 2 2.5 3 3.7 3.8 4]';
tt=[0:.1:4]';
h=0.5*g*(t+t_err).^2 + h_err;
H=[0.5*t.^2 ones(size(t))];
th=H\h
HH=[0.5*tt.^2 ones(size(tt))];
figure(4),clf
plot(t,h,'+',tt,HH*th)
xlabel 't'
ylabel 'h'
%% Problem 4
x_1=unifdist(-pi,pi,2,1);
y_1=sin (x_1);
y_2=cos (x_1);x_3=unifdist(1,2,2,1);
y_3=log (x_3);
I will appreciate any help as I am not very familiar with MATLAB yet \\ could sb please check what I've done so far
p.s. not sure if this is the right forum section for matlab-based questions
Last edited: