About

A flexible input shaping library utilizing nested simple moving averages (SMA).

nested-shaper is designed to smooth trajectories by transforming lower-order polynomial trajectories into higher-order polynomial trajectories.

• Explore the docs for detailed explanation.
• C++14 header-only library.
• Standard C++ library is not required.
• Static memory allocation in order to store samples.
• Average of euclidean or angle space. (WIP on RotationMatrix)
• Cumulative and Recursive implementations.
• Tests with Catch2. (BUILD_TESTING)
• Examples using ruckig and Matplot++. (BUILD_EXAMPLE)
• User customizable data types and average metrics. (See Advanced Guide section of docs)

Installation

nested-shaper has no dependencies except examples and testing(both are disabled by default). nested-shaper follows general steps of cmake configuration and build.

Option 1. System installation & cmake

mkdir -p build
cd build
cmake .. -DCMAKE_BUILD_TYPE=Release # -DBUILD_TESTING=ON -DBUILD_EXAMPLE=ON
make
# (sudo) make install # if you want system-wide installation

find_package(nested-shaper REQUIRED)
add_executable(myapp main.cpp)
target_link_libraries(myapp PRIVATE nested-shaper::nested-shaper)

Option 2. CMake FetchContents

include(FetchContent)
FetchContent_Declare(nested-shaper GIT_REPOSITORY
https://github.com/timetravelCat/nested-shaper)
FetchContent_MakeAvailable(nested-shaper)
...
target_link_libraries(myapp PRIVATE nested-shaper)

Option 3. Copy & Paste

Copy include/nested-shaper folder to your project. The only thing you need is C++14 compiler.

Tutorials

#include <nested-shaper/NestedShaper.hpp>

using namespace ns;

...
/**
 * First template argument "double" is your data type. If you need array type, you need to use NestedShaperEuclideanRecursiveArray with extra template arguments.
 * 
 * Second argument "5" means nested_shaper outputs position, velocity, acceleration, jerk, snap.
 * If you use "1", only outputs position.
 * Recommends to use "1", "3", "5", "7", "9", for better accuracy of derivatives.
 * 
 * Left template arguments defines maximum filter sizes.
 * In this example, there are total 3 nested filters.
 * 
 * Constructor requires at least one arguments. 
 * If one argument provided, the nested shaper is initialized with the given argument as initial position.
 * Or Constructor can provided with 1 + (total filter numbers). You can resize filter size(less than template parameter) dynamically.
 * 
 * Same behavior with constructor during runtime, by initialize() method is supported.
 **/
NestedShaperEuclideanRecursive<double, 5, FIRST_FILTER_SIZE, SECOND_FILTER_SIZE, THIRD_FILTER_SIZE> nested_shaper{3.0f};

/**
 * convolute accepts reference trajectory position and dt as arguments.
 * dt must be constant.
 * 
 * shaped_trajectory_point stores position, velocity, acceleration, jerk, snap.
 **/
array<double, 5> shaped_trajectory_point = nested_shaper.convolute(5.0, 0.01);

Code examples

• 1D euclidean : euclidean_step.cpp
• 1D euclidean, jerk limited trajectory as input : euclidean_trajectory.cpp
• 3D euclidean, jerk limited trajectory as input : euclidean_array.cpp
• Common alias classes : NestedShaper.hpp

Limitations

• Trajectory must be initialized with a arbitrary position and zero of its derivatives.

• nested-shaper can not initialized with a moving start trajectory point.

• Delta time must be same, during entire shaping process

• Maximum supported derivate order : 9 (See central_finite_difference_coefficients.hpp)

License

This project is licensed under the BSD-3-Clause - see the LICENSE file for details

Contact

timetravelCat - timetraveler930@gmail.com