Code for MIP (Mixed-Integer Programming)-based method for autonomous vehicles racing problem in head-to-head competition mode. Vehicle model: F1tenth race car. Base framework: a time-optimal NMPC framework, generated by ACADO Toolkit.

Eigen (used for linear algebra: matrices, vectors, etc)

Json (used for loading track information and other parameters)

MPCC (some pieces of code are borrowed for Spline fitting of track, related code copied in the folder "external")

ACADO (used for dynamicly linearizing and evaluating MPC problem, generated files and original source file are included in the folder "external")

qpOASES (used as QP solver, source codes are included in the folder "external")

GUROBI (used as a MIQP solver for MIP method, header file is included in the folder "external")

matplotlibcpp (used for plotting, header file is included in the folder "external")

We specify the prediction horizon length as arg1 and track as arg2, then we use a script for code generation:

# code generation and compilation
cd [this project]
mkdir build
cd build
cmake ..
bash ../config_param.sh arg1 arg2

# enter cmd line interface to test a specific scenario
./main

# or directly provide arguments in one cmd: 
# ./main s0 s0_LV total_steps ey0 epsi0 vx0 vy0 w0 d0 delta0 data_logging_file
./main 0. 0.1 146 0 0 1. 0 0 0 0 path_to_data_logging_file

# or execute batch tasks
bash ../exe_batch_N15_track1.sh