I develop this package every weekend.

Work in progress

90%

Install Requirements

$ git clone https://github.com/yukkysaito/amcl_3d.git <catkin_ws/src path>
$ rosdep install --from-paths <amcl_3d path> -y

TODO

add prediction model (particle filter)

Done

design interface (particle filter)
make ros package template
add basic resample function (particle filter)
add init function(particle filter)
build check (travis ci)
resampling timing based on ESS (effective sample size)
add simple likelihood model for lidar measurement (particle filter)
implement argumented mcl
implement kld sampling

Reference

English Reference

General

http://www.probabilistic-robotics.org/

Effective Sampling Size

http://www.cns.nyu.edu/~eorhan/notes/particle-filtering.pdf

Avoidance Numerical Underflow

http://www.maths.lu.se/fileadmin/maths/forskning_research/InferPartObsProcess/particlemethods.pdf

Japanese Reference