Code files

slam.zip
20 Kb

To run SLAM with the Kalman Filter, combine this code with the code in kf.zip (provided under Kalman Filter Localization). Includes:

  • kfslam_grid_init.m - the main script, a SLAM version of kf.m.
  • showslam.m - script to plot the path and beacon estimates after kfslam_grid_init has been run
  • several other functions called by these scripts.

 

SLAM Results

The error stats considered here are:

  • Cross-Track Error (XTE) -- the component of error orthogonal to the true robot path
  • Along-Track Error (ATE) -- the component of error tangential to the path

The errors.m script also computes

  • Cartesian error, based on the Cartesian distance between the true position and the estimate at each point in time.

Data Set A1

  

Error Statistics:

 

XTE

ATE

Mean Absolute

2.1375m

2.6191m

Maximum

5.4599m

7.5409m

Standard Deviation

1.3013m

1.3228m

Data Set A2

  

Error Statistics:

 

XTE

ATE

Mean Absolute

1.7223m

1.8231m

Maximum

12.3644m

12.7233m

Standard Deviation

1.0758m

1.0632m

Data Set B1

  

Error Statistics:

 

XTE

ATE

Mean Absolute

--m

--m

Maximum

--m

--m

Standard Deviation

--m

--m

Data Set B2

  

Error Statistics:

 

XTE

ATE

Mean Absolute

--m

--m

Maximum

--m

--m

Standard Deviation

--m

--m

Data Set C1

  

Error Statistics:

 

XTE

ATE

Mean Absolute

0.4768m

0.3743m

Maximum

2.7699m

1.8624m

Standard Deviation

0.3469m

0.3072m

Data Set C2

  

Error Statistics:

 

XTE

ATE

Mean Absolute

0.5981m

0.6573m

Maximum

1.5984m

2.0863m

Standard Deviation

0.3898m

0.4737m

Data Set C3

  

Error Statistics:

 

XTE

ATE

Mean Absolute

--m

--m

Maximum

--m

--m

Standard Deviation

--m

--m