Particle Filter algorithm implementation and testing
Particle Filter is used to estimate probability distribution of current state in Hidden Markov Models where state space is continuous. Density of discrete particles approximates probability distribution which is continuous function.
pfilter module gives universal Particle Filter implementation
robot module uses particle filter to localize robot named Paul in the maze. Paul is completely blind and the only two thing the knows on each step are:
-
Whether he ran into a wall or not.
-
Distance to each beacon.
USAGE: python robot.py [PARTICLE_COUNT [BEACON_COUNT [SIGMA [MAZE_FILE]]]