Transcript Lecture 9

ASEN 5070: Statistical Orbit Determination I
Fall 2015
Professor Brandon A. Jones
Lecture 9: Batch Estimator and Weighted LS
University of Colorado
Boulder

Homework 3 – Due September 18

Lecture Quiz 3 Due Friday by 5pm

Future Lectures
◦ Lecture 10 – Monday 9/14 @ 4pm
◦ Lecture 11 – Monday 9/21 @ 9am
◦ Lecture 12 – Monday 9/21 @ 4pm
University of Colorado
Boulder
2
Batch Estimator
University of Colorado
Boulder
3
University of Colorado
Boulder
4


Straightforward way to estimate the state at a
time that matches the observations
What about when the observations cover
multiple points in time?
University of Colorado
Boulder
5

What can we do to estimate the state x0 when we
have observations at multiple points in time?
◦ What tool(s) do we have
available to alter the
formulation?
◦ Given result from above,
how might we alter the
formulation to use a single
relationship of the form:
University of Colorado
Boulder
6
University of Colorado
Boulder
7
University of Colorado
Boulder
8

Process all observations over a given time
span in a single batch
◦ The alternative sequential methods will be
discussed later

What are the shortcomings of such a
formulation?
University of Colorado
Boulder
9
University of Colorado
Boulder
10
University of Colorado
Boulder
11
University of Colorado
Boulder
12

Process all observations over a given time
span in a single batch
University of Colorado
Boulder
13

No weighting of observations
◦ How do we account for different sensors with
different accuracies?

No incorporation of previous information
◦ Known a a priori state information
◦ How do we include this in the filter?
University of Colorado
Boulder
14
Weighted Least Squares Estimation
University of Colorado
Boulder
15

For each yi, we have some weight wi
University of Colorado
Boulder
16


Consider the case with two observations
(m=2)
If w2 > w1, which εi will have a larger
influence on J(x) ? Why?
University of Colorado
Boulder
17
University of Colorado
Boulder
18

For the weighted LS estimator:

How do we find W ?
University of Colorado
Boulder
19
University of Colorado
Boulder
20
University of Colorado
Boulder
21
Weighted Least Squares w/ A Priori
University of Colorado
Boulder
22

A priori
◦ Relating to or denoting reasoning or knowledge
that proceeds from theoretical deduction rather
than from observation or experience

We have:
University of Colorado
Boulder
23

As you will show in the homework:
University of Colorado
Boulder
24
University of Colorado
Boulder
25