LearnRoboticsCpp

This repository contains my implementations of classical robotics algorithms in C++. Inspiration drawn from PythonRobotics and CppRobotics. The CppRobotics repo was very good, but it used OpenCV to plot 2D graphs. I found the process of converting 2D points to pixel coordinates in OpenCV very tedious and it seems like a bit of a hack. This repo uses gnuplot-iostream instead for plotting which makes much prettier graphs than OpenCV and allows for us to easily make 3D plots.

Table of Contents

• Requirements
• Dependencies Installation
• Build
• Localization
  • Extended Kalman Filter
  • Unscented Kalman Filter
  • Particle Filter
• Path Planning
  • Dijkstra
  • AStar
  • RRT
  • RRTstar
  • PRM
  • Potential Field
  • Quintic Polynomial
  • Cubic Spline
  • DWA
  • Model Predictive Trajectory Generator
  • State Lattice Planner
• Path Tracking
  • Move to Pose
  • Stanely Control
  • Model Predictive Control

Requirments

Tested on Ubuntu 18.04

• cmake
• opencv 3.3 (for KD tree in PRM)
• Eigen 3
• Boost 1.4 (for gnuplot-iostream)
• gnuplot
• ipoptd (this one is a pickle, install tips borrowed from Udacity )
• cppad

Dependencies Installation

1. apt installs

sudo apt update
sudo apt install build-essential
sudo apt install cmake
sudo apt install gnuplot
sudo apt install libboost-all-dev
sudo apt install libopencv-dev python3-opencv
sudo apt install libeigen3-dev
sudo apt install cppad

2. ipoptd install

sudo apt install gfortran
sudo apt install unzip

You probably want to cd into another directory like Downloads for this folder

wget https://www.coin-or.org/download/source/Ipopt/Ipopt-3.12.7.zip && unzip Ipopt-3.12.7.zip && rm Ipopt-3.12.7.zip

Call install_ipopt.sh in this repo with sudo and the source directory as the first argument (e.g. sudo bash install_ipopt.sh ~/Ipopt-3.12.7)

3. Xming (for Windows Subsystem Linux users only) Need to install xming for gnuplot visualization: https://sourceforge.net/projects/xming/. Set display to gnuplot with: $ export DISPLAY=:0 gnuplot.

Build

$ git clone https://github.com/CtfChan/LearnRoboticsCpp.git
$ mkdir build
$ cd build
$ cmake ../
$ make -j4

Path Planning

DWA

Dijkstra

A*

PRM

RRT

RRTStar

Potential Field

Quintic Polynomial

Cubic Spline

Model Predictive Trajectory Generation

State Lattice Planner

Path Tracking

Move To Pose

Stanley Control

Model Predictive Control

Localization

Extended Kalman Filter

Unscented Kalman Filter

We can see error ellipse in this demo is a much better approximation of the true distribution. EKF can be biased and inconsistent.

Particle Filter