https://github.com/juliarobotics/approxmanifoldproducts.jl

Approximate the product between infinite functional objects on a manifold -- i.e. belief products
https://github.com/juliarobotics/approxmanifoldproducts.jl

Last synced: 2 months ago
JSON representation

Approximate the product between infinite functional objects on a manifold -- i.e. belief products

• Host: GitHub
• URL: https://github.com/juliarobotics/approxmanifoldproducts.jl
• Owner: JuliaRobotics
• License: mit
• Created: 2018-10-01T17:16:31.000Z (almost 8 years ago)
• Default Branch: develop
• Last Pushed: 2025-05-05T00:41:52.000Z (about 1 year ago)
• Last Synced: 2025-07-09T05:48:38.513Z (12 months ago)
• Language: Julia
• Size: 655 KB
• Stars: 12
• Watchers: 4
• Forks: 4
• Open Issues: 12
• Metadata Files:
  • Readme: README.md
  • Changelog: NEWS.md
  • License: LICENSE

Awesome Lists containing this project

README

          
# ApproxManifoldProducts.jl

[![CI](https://github.com/JuliaRobotics/ApproxManifoldProducts.jl/actions/workflows/ci.yml/badge.svg)](https://github.com/JuliaRobotics/ApproxManifoldProducts.jl/actions/workflows/ci.yml)

[![codecov.io](https://codecov.io/github/JuliaRobotics/ApproxManifoldProducts.jl/coverage.svg?branch=master)](https://codecov.io/github/JuliaRobotics/ApproxManifoldProducts.jl?branch=master)

# Introduction

Approximate the product between infinite functional objects on a manifold -- i.e. belief products.  ApproxManifoldProducts (AMP) is intended for use with the [Caesar.jl](http://www.github.com/JuliaRobotics/Caesar.jl) framework.  AMP is intended for development and testing of various kernel embedding methods for approximating functional products.

See [Caesar documentation](http://www.juliarobotics.org/Caesar.jl/latest/) for more details.

# Installation

For Julia 0.7 and above press ']' in the Julia REPL:

```julia

pkg> add ApproxManifoldProducts

```

# Current Supported (Mixed) Manifolds

**NOTE** Current work on v0.3.x is transitioning to JuliaManifolds for some of the Manifold abstractions.  This package AMP then adds non-parametric probabilistic computations on those and user manifolds.

The following on-manifold function approximations:

- Euclidean (2DOF),

- S1/Circular (1DOF), or SO(2) equivalent.

- SE(2)

- S2 (not implemented yet)

Work in progress on so-called 'subgroup' mixed-manifolds, where DOFs are bunched together for particular manifolds:

- SO(3) / Quaternion,

- Plucker coordinates (SP(3)) for rigid transforms,

- SE(3) for rigid transforms.

> Implementation of multivariate methodology requires code to consider "double loops" that iterate over sub-groups, and then within each sub-group as required.  The supported list above can be implemented with "single loop" over all DOFs.

Any suggestions or issues welcome.