Vehicle Path Follow Simulation in Ubuntu

General description

• Create a path following simulation for vehicle using RK4, PID(or LQR) and Bicycle Model in C++
  • Minimalism code style, no 3rd parties libraries (except Eigen)

User can choose any modes :

• Controller (P/PID/LQR)
• Discrete Propagation(Forward Euler/RK4 aka Runge Kutta Method)
• Vehicle Dynamics (Naive/Advanced Bicycle model, Pacejka Tire force model will be updated later)
• Waypoint Generator (P2P/Sinusoidal/Cubic/Zigzag)
• Using/Not Using Low Pass Filter in PID

Please modify which modes you want in the "config" Object. Of course, for different scenario all gains need to be tuned again. For LQR, number of maximum iterations and cost weights Q and R need to be tuned properly. For LQR, the discrete time algebraic Riccati equation (DARE) is solved iteratively.

Requirement packages

• g++ in Linux Ubuntu

sudo apt-get install g++

• Gnuplot

sudo apt-get install gnuplot

• Eigen (OPTIONAL, only installed if you want to work with LQR later)

sudo apt install libeigen3-dev

Usage

• Just compile normally as regular C++ code:

  g++  -std=c++14 -Wall -O3 -g VehiclePathFollowSim.cpp -o VehiclePathFollowSim
  ./VehiclePathFollowSim

• Please note that to use LQR, you need to define the EIGEN_LIB, just simply comment/uncomment this line:

#define EIGEN_LIB 1
//Comment means no use, uncomment means eigen library will be used

• To change the controller type, in the Config class , modify:

  const CONTROLLER_ALG controller = LQR_CONTROL;
  //Note it can also be P_CONTROL or PID_CONTROL

• To change the waypoint trajectory, either toggle these 2 lines:

    const bool useZigZagWay = false; //or Sample/P2P
    const bool useSampleWay = true; //or Zigzag/P2P

• To change different propagation method (RK4/Euler) or Vehicle Dynamics Model(Naive/Advanced), change the enum :

  const PROPAGATOR_MODE propagator = RK4_NAIVE_DYNAMICS; 
  //Note it can be  RK4_NAIVE_DYNAMICS , EULER_NAIVE_DYNAMICS 
	// RK4_ADV_DYNAMICS , EULER_ADV_DYNAMICS 

• To use Low pass filter, change the value of beta (from 0 to 1, 0 means no use):

    const float beta = 0.9; // for low pass filter

Notes

This is still under developing progress. Kalman filter/Particle Filter will be added soon. You can check this video that I performed : https://www.youtube.com/watch?v=G9c8e5KYjUQ