I am doing a project on Indoor Positioning. I have used NS2 and a matlab GUI named Senelex to find the original and estimated positions of nodes. Now I want to use a Kalman filter to estimate the position of the node/target.
For example:
If I consider the velocity of the target as constant, how can I use the determined values of estimated or original position of nodes to provide as input to a Kalman filter.
The following are the estimated position and original position of nodes.
ori = [6.62650602409639 194.444444444445;
6.62650602409639 10.6837606837607;
192.168674698795 7.83475783475797;
192.168674698795 191.595441595442;
70.4819277108434 171.652421652422;
129.518072289157 168.803418803419;
24.6987951807229 144.586894586895;
42.7710843373494 79.0598290598291];
est = [6.62650602409639 194.444444444445;
6.62650602409639 10.6837606837607;
192.168674698795 7.83475783475797;
192.168674698795 191.595441595442;
70.7600705547484 171.847603055024;
129.443055817301 168.734648868329;
25.01956026761 144.890243978875;
42.6058125534278 79.1446327727804];
How can I use these as inputs to a Kalman filter and estimate the target using a Kalman filter?
I’d suggest by starting with the relatively straight forward tutorial on Kalman filters: An Introduction to the Kalman Filter. The Kalman Filter site has fair number of good resources, including links to a Matlab toolbox. The Kalman filter implementation is not a very complex program, once you have the equations.
Here are some links to some Java versions of Kalman filters:
I’m sure there is source code available for most programming languages. These examples are ones with which I’ve worked.