Blast545/cppl1_q12020

cpp_Q12020

cpp

C++ Isometry Library Challenge Source repository

CI Status

• master CI
• devel CI

Description

This project contains the basic set of files needed to kick-start the C++ Isometry Library challenge for C++ L1 padawans.

The folder contains both the actual project and a development container.

The goal of this challenge is to write a small library that enables us to determine transformations between coordinate frames in a 3D space, by defining vectors, rotation matrices and isometries to performn the transformations using regular algebrac expressions.

The devlopment container provides a development environment that has all the tools needed for the build process pre-installed. Using this container is optional but recommended, since it provides all the tools needed to build and process the project, with the correct versions needed for it.

Project organization

The current folder has the following subfolders.

• course: Source code of the challenge.
  • include: Header files.
  • src: Source files.
  • test: Testing files.
    • gtest: GTest library source files. Do to alter these files.
    • src: Test files to check the generated code.
• docker: docker related files.

Problem statement

Write a small library to work with isometries. Those transformations should be able to correctly express:

• 3D composed translations and rotations.
  • From only translation vectors.
  • From only (Euler) rotation angles.
  • From composed translation vectors and rotation angles.
  • From pre-made useful constants, such as unit vector translation in x axis, etc.
• Composition of transformations.
  • Direct of a 3D coordinate in space A to space B.
  • Inverse of a 3D coordinate in space B to space A.
• IO streaming for ease of use and debugging.
• double support. → There is no need for any other data type.

The interface required for your library is defined through the test files that you will find in the test/src folder.

You will need to develop the following code abstractions to satisfy the tests:

• Vector3: Three element vector, to designate an (x, y, z) coordinate in a coordinate frame.
• Matrix3: 3x3 elements matrix, to write rotation matrices.
• Isometry. A homogeneus-matrix abstraction.

We encourage you to consider using the following namespaces:

• ekumen::math
• ekumen::math::test

Bibliography

• Planning algorithms, Steven LaValle, Cambridge University. Chapter 3.
• Homogeneous Transformations notes.

Requirements

• Build an isometry tranformations matrix from from:
  • Translation vector.
  • Euler rotation angles.
  • Translation vector and Euler angles.
• Compose transformations in mathematical and semantic fashion.
• Let T be a transformation from space A to space B, and M be the transformation from space B to space A. Provide a mathematical and semantic way to obtain M from T.
• Given a point P in space A, transform it to space B with the isometry T (A → B).
• Given a point Q in space B, transform it to space A with the isometry T (A → B).
• Obtain the translation and rotation (matrix) parts from an isometry.
• Serialize to std::string a vector, a matrix and the isometry.
• Compare equality (to the type epsilon extent) of a vector, matrix and an isometry.
• Satisfy the provided unit tests of the above features.

Installation

1. Fork and clone this repository.

2. Build docker image and run container, visit docker readme for instructions.

3. You can now begin working on the code.

To build and test your code:

bash
cd {REPO_PATH}/course
mkdir build
cd build
cmake ..
make
ctest