Expand description
Monte Carlo localization of a 2D LiDAR in a known occupancy grid.
- Motion: every particle applies the odometry step with noise that grows with the distance and rotation travelled (odometry motion model).
- Measurement: the likelihood-field model scores a subsampled scan by the distance from each beam end point to the nearest occupied cell, mixed with a uniform term for unexplained returns.
- Global initialization: the first scan after
LidarMcl::initialize_globalscoresglobal_oversamplingtimes as many uniform poses as there are particles and keeps the best by weight. - Recovery: augmented MCL tracks short- and long-term averages of the measurement likelihood and injects new particles when the short-term average drops; sensor resetting also injects while the scan fit at the particles stays poor. Each injected particle is the best-scoring of a few uniform free-space draws, so the filter recovers from a kidnapping without needing a lucky sample. A small fraction of scored particles is injected on every update, so a filter that converged on a look-alike place (long uniform corridors) keeps probing for the true one.
- Resampling: weights accumulate across updates and the particles are resampled (low variance) only when the effective sample size halves.
Structsยง
- Lidar
Mcl - Particle-filter localization against an
OccupancyGrid. - Lidar
MclConfig - Tuning parameters for
LidarMcl. - Particle
- A weighted pose hypothesis.