Monte carlo locatization with robot Mikes - Tomas Kubla

Task

We had to implement MCL-Algorithm (Sebastian THRUN, Wolfram BURGARD, Dieter FOX: PROBABILISTIC ROBOTICS, 1999-2000, page 200)

Creating tag model

We were putting robot to many positions in arena and we were logging:

• X position
• Y position
• azimuth
• list of seen tags

Algorith

 int init_mcl(){
   ...
   for(int i = 0; i< HYPO_COUNT; i++){
           hypo[0][i].x = hypo[1][i].x = rand() % 280;
           hypo[0][i].y = hypo[1][i].y = rand() % 280;
           hypo[0][i].alpha = hypo[1][i].alpha = rand() % 360;
           hypo[0][i].w = hypo[1][i].w = 0.3;
           ...
   }
   ...
 }

Source Code

Source code is located on [1]

Some of support scripts and data are here: data.zip