Crate multisensor_lmb_filters_rs 

Prak - Multi-object tracking library

Rust implementation of multi-object tracking algorithms based on Labelled Multi-Bernoulli (LMB) filters and their variants.

Features

• Single-sensor LMB and LMBM filters
• Multi-sensor fusion algorithms (PU-LMB, IC-LMB, GA-LMB, AA-LMB)
• Multiple data association methods (LBP, Gibbs, Murty’s algorithm)

Modules

• lmb - LMB tracking algorithms and types
• components - Shared algorithms: prediction, update
• association - Data association: likelihood computation, matrix building
• common - Low-level utilities

Example

use multisensor_lmb_filters_rs::lmb::{Filter, LmbFilter, MotionModel, SensorModel, BirthModel, BirthLocation, AssociationConfig};
use nalgebra::{DVector, DMatrix};

// Create filter configuration
let motion = MotionModel::constant_velocity_2d(1.0, 0.1, 0.99);
let sensor = SensorModel::position_sensor_2d(1.0, 0.9, 10.0, 100.0);

// Define a birth location
let birth_loc = BirthLocation::new(
    0,
    DVector::from_vec(vec![0.0, 0.0, 0.0, 0.0]),
    DMatrix::identity(4, 4) * 100.0,
);
let birth = BirthModel::new(vec![birth_loc], 0.1, 0.01);
let association = AssociationConfig::default();

// Create filter
let mut filter = LmbFilter::new(motion, sensor, birth, association);

// Process measurements
let mut rng = rand::thread_rng();
let measurements = vec![DVector::from_vec(vec![1.0, 2.0])];
let estimate = filter.step(&mut rng, &measurements, 0).unwrap();

Re-exports

pub use lmb::AssociationConfig;

pub use lmb::BirthLocation;

pub use lmb::BirthModel;

pub use lmb::EstimatedTrack;

pub use lmb::FilterOutput;

pub use lmb::FilterParams;

pub use lmb::FilterThresholds;

pub use lmb::GaussianComponent;

pub use lmb::LmbmConfig;

pub use lmb::LmbmHypothesis;

pub use lmb::MotionModel;

pub use lmb::MultisensorConfig;

pub use lmb::SensorModel;

pub use lmb::SensorVariant;

pub use lmb::StateEstimate;

pub use lmb::Track;

pub use lmb::TrackLabel;

pub use lmb::Trajectory;

pub use lmb::AssociationError;

pub use lmb::FilterError;

pub use lmb::Associator;

pub use lmb::Filter;

pub use lmb::Merger;

pub use lmb::Updater;

pub use lmb::GibbsAssociator;

pub use lmb::LbpAssociator;

pub use lmb::MurtyAssociator;

pub use lmb::HardAssignmentUpdater;

pub use lmb::MarginalUpdater;

pub use lmb::LmbFilter;

pub use lmb::LmbmFilter;

pub use lmb::AaLmbFilter;

pub use lmb::ArithmeticAverageMerger;

pub use lmb::GaLmbFilter;

pub use lmb::GeometricAverageMerger;

pub use lmb::IcLmbFilter;

pub use lmb::IteratedCorrectorMerger;

pub use lmb::MultisensorLmbFilter;

pub use lmb::MultisensorLmbmFilter;

pub use lmb::MultisensorMeasurements;

pub use lmb::ParallelUpdateMerger;

pub use lmb::PuLmbFilter;

pub use lmb::MultisensorAssociationResult;

pub use lmb::MultisensorAssociator;

pub use lmb::MultisensorGibbsAssociator;

Modules

association

Data association algorithms and utilities Data association algorithms and likelihood computation

bench_utils

Benchmark utilities (scenario loading, filter factory) Benchmark utilities shared between Criterion benchmarks and the benchmark_single binary.

common

Low-level utilities (linear algebra, RNG, constants) Common utilities and shared components for tracking algorithms.

components

Shared tracking components (prediction, update) Core algorithmic components

lmb

LMB (Labeled Multi-Bernoulli) tracking algorithms

Constants

VERSION

Library version