This code is a part of https://github.com/ethz-asl/mav_control_rw.git.

Author: Oskar Ljungqvist Date: 2017-11-23 Description:

This folder contains code to automatically generate an MPC controller using ACADO toolkit.

The file nmpc_solver_setup.cpp is what defines the MPC controller and it is this you should change in order to change the controller. Then build/install it using the explanation below. As an example it easy to change the prediction horizon if that is prefered.

ACADO toolkit needs to be properly installed. Follow the instructions here: http://acado.github.io/install_linux.html

• In your ~/.bashrc add the line:source [pathToAcado]/ACADOtoolkit/build/acado_env.sh. Remember to source ~/.bashrc,

source ~/.bashrc

• Build/Install. From this folder do:

cd [yourDirectory]/make_acado_solver
mkdir build
cd build
cmake ..
make

• The executable file is now placed in the folder ../solver. Execute it by running the following commands:

cd ../solver
./nmpc_solver_setup

Then the controller is generated on solver/OCPexport folder.

• Copy your controller to your project, i.e.

cp -r ./OCPexport/ ../test/solver/

• There is an example of how to use the generated controller on test folder.

cd ../test
mkdir build
cd build
cmake ..
make
./main

• 下载安装acado, http://acado.github.io/install_linux.html.

• 增加环境变量，在 ~/.bashrc 增加一句 source [你的路径]/ACADOtoolkit/build/acado_env.sh 加完这一句后，记得要 source 一下

source ~/.bashrc

• 构建求解器，在nmpc_solver_setup.cpp上构建你的求解器

• 生成求解器

cd [你的路径]/make_acado_solver
mkdir -p build
cd build
cmake ..
make 
cd ../solver
./nmpc_solver_setup

然后会在solver/OCPexport文件里生成你的解析器了

• 将生成的控制器拷贝到测试文件夹test中

cp -r ./OCPexport/ ../test/solver/

• 编译运行例程

cd ../test
mkdir build
cd build
cmake ..
make
./main

每次改动求解器后，要记得删除./test/solver/OCPexport，将新的./solver/OCPexport文件夹替换原来的文件夹