/train-platform

Primary LanguageJupyter Notebook

Introduction

Train-Platform is my graduation project which is intended both as a theory reference for scientists to understand related work and as a experiment tool for engineers to practice their controller design.

Features

This project initially tends to study cooperative control theory and methods. Besides this requirement, I spent much time for extensive functions and now this platform includes:

  • dynamics model
  • controller design
  • computer simulation
  • visualization
  • documentation

Configuration

  1. Gurobi

  2. spdlog

After you configure these third-party, you should change address information in makefile.

Getting Started

  1. dp

    make dp_ref
./dp_ref

  2. leader mpc

    make leader
./leader

  3. follow mpc

    make convoy
./convoy

  4. data-driven mpc

    make data_driven_leader
./data_driven_leader

File Structrue

train-platform
├─ README.md
├─ db
├─ examples
│  └─ demo
├─ include
├─ logs
├─ makefile
├─ src
│  ├─ calculate
│  │  └─ follow_MPC.cpp
│  ├─ communication
│  ├─ config
│  │  └─ logger.cpp
│  ├─ control
│  │  ├─ follow_controller.cpp
│  │  └─ predictor.cpp
│  ├─ environment
│  │  ├─ read_speed_limit.cpp
│  │  └─ read_speed_max.cpp
│  ├─ model
│  │  └─ dynamic_model.cpp
│  └─ programming
├─ user
│  ├─ analysis
│  ├─ result
│  └─ visualization
└─ utils
   └─ exception

Model

continuous form

$$ \dot{x} = v \tag{1} $$

$$ Ma = -A - Bv-T_fCv^2-F_e+u \tag{2} $$

discrete form

$$ v_{i+1}^2 - v_{i}^2 = 2a_is \tag{3} $$

$$ Ma_i = -A-Bv_i-T_fCv_i^2-F_e+u_i \tag{4} $$

Maintainers

author zhihu