Box Particle Filter

Matlab code for the Box Particle filtering project. Implementation of Interval class and simulation applications.

Directories

BPF_ND:

2D Box Particle Filter implementation.

• BPF.m: N-dimenional Box Particle Filter simulation.
• measurementUpdate_ND.m: Measurement update for N-dimensional box particle filtering.
• stateUpdate.m: State Update for N-dimensional box particle filtering.
• initBoxesArray.m: Initialise Box cell array. Used by stateUpdate function to perform resampling.
• findIndexes.m: Find which boxes in a box array intersect a given input box.

BPF2D:

2D Box Particle Filter implementation.

• BPF2D.m: 2D Box Particle Filter simulation.
• measurementUpdate.m: Measurement update for 2D box particle filtering.
• stateUpdate.m: State Update for 2D box particle filtering.
• initBoxesArray.m: Initialise Box cell array. Used by stateUpdate function to perform resampling.
• findIndexes.m: Find which boxes in a box array intersect a given input box.

Obs.: the BPF2D function performs a full simulation. If something else must be done after each iteration, a loop calling the functions measurementUpdate and stateUpdate must be programmed manually (see BPF2D for the implemetation).

BPF2D_fixed:

2D Box Particle Filter implementation for a fixed rectangular room [Not used anymore].

• BPF2D_fixed.m: 2D Box Particle Filter simulation.
• measurementUpdate_fixed.m: Measurement update for 2D box particle filtering.
• stateUpdate_fixed.m: State Update for 2D box particle filtering.
• initBoxesArray.m: Initialise Box cell array. Used by stateUpdate function to perform resampling.
• findIndexes_fixed.m: Find which boxes in a box array intersect a given input box.

ConventionalParticleFilter:

Conventional particle filter implementation.

SLAM:

Simple implementation of the SLAM algorithm using the BPF2D as estimator engine.

Interval:

Implementation of intervals.

Mist:

Plot helper functions.

• plotBoxes.m - plot of each box in a box array
• plotBoxesColor.m - plot of each box in a box array with colouring - representing each box's weight
• plotDistance.m - plot of distance between each point in two vectors

Simulations:

Multiple example environments defined to test the SLAM algorithm and the Box Particle Filter implementations