LiU Racetrack 2018

About this project

Introduction and background to this project

Introduction

Development of strategies and algorithms for automatic control of vehicles is a hot topic in both research and industry. The department for Automatic Control has created a framework consisting of RC-vehicles, IR-cameras and tracking algorithms to be used in various projects, both for research and CDIO.

Background

This project has been an ongoing project for several years where multiple CDIO-projects as well as summer-projects have been conducted. Previous years have developed automatic control-algorithms for driving through a racetrack-course for several vehicles, both in forward and backward movement. They have also developed algorithms for overtaking of slower vehicles and an associated motion planer.

Project Scope

This years project had the following objective:

  • Modularize and improve the existing system using ROS (Robot Operating System).
  • Develop an automatically controlled LEGO-truck with associated motion planner, state observer and controller for trajectory following.

Video

Results

LEGO-truck: Lattice Planner

Motion Primitive Generation
  • A set of feasible trajectories (motion primitives) are computed that moves the vehicle from and to discrete vehicle states.
  • The set of motion primitives are computed offline using numerical optimal control.
Online Planning
  • The motion primitives are combined online to generate a reference trajectory from the vehicle's initial state to a desired goal state.
  • The optimal sequence of motion primitives that connects the start and the goal is found using A*-graph search where an admissible heuristic function is used to guarantee resolution optimality.

LEGO-truck: Path Following Controller

LQ-Controller
  • A path following controller has be designed using Linear Quadratic (LQ) control.
  • Different feedback gains depending on the direction of motion.
  • The lateral deviation from the reference trajectory is less than 5 cm when driving backwards around the figure eight.
Positioning System
  • The Qualisys positioning system in Stora Visionen is used for global positioning.
  • On-board angular sensors are used to measure the relative angles of the truck and trailer.

Migration to ROS

  • A thorough analysis of the old system was made in order to identify possible improvements and modularizations.
  • An attempt was made to map the old system to ROS, but the lack of time made the task impossible to complete.
  • Several GUIs was dveloped using ROS that can be used in other parts of the system, including the LEGO-truck.

Documentation

The Project Members

The Amazing Group!