ParticleFilters.md (1021B)
1 # Particle Filters 2 3 **Source:** Probabilistic Robotics 4 5 **Chapter:** 4.2 6 7 **Definition:** Particle filters are an approach to localization where many particles are placed in an environment where each particle has the same sensing abilities as the robot. At each sense step we then compute importance weights for each particle and sample on the basis of this value. 8 9 ## Steps 10 11 1. Initialize random, valid particles 12 2. Main Loop 13 1. Move each particle based on the robot movement command 14 2. Compute the importance weight for each particle 15 1. This is proportional to how likely the particle measurement is given the robot's measurement. 16 3. Sample particles proportional to their importance weight 17 1. This involves sampling particles with replacement to survive to the next step. This ensures we have a consistent number of particles from turn to turn, and that they are decent representations of the robot's true state. 18 4. Fuzz (apply pertubations to surviving particles to induce diversity)