latest algorithms to improve the gps signal accuracy

i want to use kalman, neural network and neuro fuzzyto improve gps signal accuracy.
please help me to start my work..

oh come on!!!
atleast choose your titles carefully. At first it occured to me you've developed some new algorith to improve gps signal accuracy.
Anyways, I can help you up with kalman a bit. for kalman, you need two ways to measure the gps location, then you can decide which one to use for prediction and which one for correction
Read a few papers about kalman, then we ca discuss

Hi..

"KALMAN FILTERING APPLICATION WITH MATLAB" by Rachel Kleinbauer, gave me the complete theoretical knowledge with matlab code. chapter 6 follows like this.........
...................................................................................................................................................................
6.1 Extraction of the data

The first step of the program is reading of the required data from text files that the user has to prepare before using ht program. How the user has to format the text files is shown in the folder 'format'.

In this way the following data is read in:

state vector* x_sym
initial value of the state vector x0
its covariance matrix P0
differential equation of the dynamic model * f
covariance matrix of the dynamic noise Q
equation of the observation model h
observations l
their covariance matrix R
the time step of observations t
..................................................................................................................................................................

now my question is that how to do it?

1. Shall i buy a new GPS receiver to collect the text file?
2. Then is it possible to arrange the file in the above mentioned order?

Thank you..

well, if you just want to do analysis first, i dont think you need to buy a gps reciever for that.
You can define your own vectors. you dont even need to go for a text file in that case. Take some random starting co-cordinate (may be you can start with your current location). then generate state vector by using any random line equation. say Y= ut + 1/2 at^2. you can add some white noise or pink noise which ever you like. Define some sampling rate(say 2ms).
Define other way of measuring (eg. you read your velocity as well). so your ΔY = vΔt, here Δt is sampling time and v is instantaneous velocity. and you can very well generate a random velocity vector in Matlab.
phew!!! you have two equations. all the required readings as well. Rest you can calculate using these values.
Apply Kalman filter, tune it according to your need.

Hi..

I got the idea of kalman filter in some other dimension from the above said tutorial.
so i am not able to understand your explanation.
will you give me any link to know the basics of kalman filter? (for understanding your suggestion)

- - - Updated - - -

Hi..

I got the idea of kalman filter in some other dimension from the above said tutorial.
so i am not able to understand your explanation.
will you give me any link to know the basics of kalman filter? (for understanding your suggestion)

yeah I think you should better be starting with the basics. Like how it works and where can you use it
May be this link will help you out. **broken link removed**
This one is very much similar to your end application

Thank you..