Skip to content

examples.md

Latest commit

 

History

History
116 lines (80 loc) · 3.17 KB

examples.md

File metadata and controls

116 lines (80 loc) · 3.17 KB

Examples

The source repository includes executable examples in examples/basic/.

Run examples from the repository root after installing PyRecEst or after installing a development checkout.

Basic Examples

Gaussian Multiplication

examples/basic/gaussian_multiplication.py multiplies several two-dimensional Gaussian distributions and checks the result against the closed-form information representation.

python examples/basic/gaussian_multiplication.py

Kalman Filter

examples/basic/kalman_filter.py runs a one-dimensional constant-velocity Kalman filter.

python examples/basic/kalman_filter.py

Kalman Filter With Model Objects

examples/basic/kalman_filter_with_models.py runs the same one-dimensional constant-velocity Kalman filter as kalman_filter.py, but defines reusable linear-Gaussian transition and measurement model objects and passes them to predict_model() and update_model().

python examples/basic/kalman_filter_with_models.py

Unscented Kalman Filter With Model Objects

examples/basic/ukf_with_models.py demonstrates the additive-noise nonlinear model-object API with an unscented Kalman filter. The transition model stores the state propagation function and process-noise covariance; the measurement model stores the measurement function and measurement-noise covariance.

python examples/basic/ukf_with_models.py

This example follows the current backend limitations of UnscentedKalmanFilter.

Particle Filter With Model Objects

examples/basic/particle_filter_with_models.py demonstrates a particle-filter loop with a sampleable transition model and a likelihood-based measurement model.

python examples/basic/particle_filter_with_models.py

Multi-Target Tracking

examples/basic/multi_target_tracking.py runs a small linear/Gaussian multi-Bernoulli-tracker scenario with two labeled targets, missed detections, and clutter measurements.

python examples/basic/multi_target_tracking.py

This example currently requires the NumPy backend.

von Mises-Fisher Multiplication

examples/basic/von_mises_fisher_multiplication.py multiplies two von Mises-Fisher distributions on the unit sphere and verifies the analytic product relation.

python examples/basic/von_mises_fisher_multiplication.py

Custom Distribution Protocol

examples/basic/custom_distribution_protocol.py shows a small user-defined scalar distribution that satisfies the public dimension protocols without inheriting from a PyRecEst distribution base class.

python examples/basic/custom_distribution_protocol.py

Custom Filter Protocol

examples/basic/custom_filter_protocol.py shows a small user-defined recursive filter that satisfies the public dimension protocol and follows common PyRecEst filter naming conventions.

python examples/basic/custom_filter_protocol.py

Backend Selection

Select a non-default backend by setting PYRECEST_BACKEND before running the script. For example, on a bash-compatible shell:

PYRECEST_BACKEND=pytorch python examples/basic/kalman_filter.py

Install the matching optional dependency extra before using a non-default backend.