/Motion-Planning-Carla-Simulator

Repository designed to store and disseminate the Python scripts developed by me for an Institutional Research (CNPq - PIBITI) aimed at motion planning of autonomous vehicles in the Carla simulator

Primary LanguagePythonMIT LicenseMIT

Motion Planning on CARLA Simulator

Objective

Development of a motion planning solution implemented in Python using the Carla Simulator.

Usage

Configuration

  • The following python scripts were developed in Python 3.7.11, through AnaConda.

  • The Operational System was Ubuntu 18.04

  • The CARLA version was 0.9.11, downloaded following the Debian installation from this installation guide.

  • Processor: Pentium G4560

  • Graphic Card: GTX 1050Ti

  • Memory: 16Gb DDR4 2400mhz

How to Run

  • Add the contents of the 'agents' folder of this repository in the '/opt/carla-simulator/PythonAPI/agents' folder in the your local CARLA.
  • Add the content of the 'examples' folder of this repository in the '/opt/carla-simulator/PythonAPI/examples' folder in the your local CARLA.
  • Install the Python packages from 'requirements.txt' with Pip. For example pip install -r requirements.txt
  • In the '/opt/carla-simulator/PythonAPI/examples' folder, run python motion_planning.py

Acceptance criteria

As for the research project

  • The student must develop a script in Python that solves the motion planning problem for autonomous vehicles in CARLA simulator.
  • The student must prepare a scientific article on motion planning and the developed solution.
  • The student must present both the script and the article to the supervising professor for appropriate corrections, adjustments and eventual approval of the activities developed during the research period.
  • The deadline of the activities is October 31, 2021.

As for the trajectory

  • The vehicle must be able to reach any point B from any other point B.
  • The trajectory and the waypoints must be selected and positioned according to the OpenDRIVE high definition map of the region.
  • The route generated to the destination must be efficient, that is, without being longer than necessary;
  • The chosen route must obey the traffic laws

As for the motion planning

  • The vehicle must travel the route obeying the current traffic laws, such as:

    • Maximum speed
    • Minimum speed
    • Lane centering
    • Lane changes
    • Lane Direction
    • Traffic lights
    • Signs

  • The vehicle must travel the route safely, avoiding accidents such as:

    • Collisions with other vehicles (front, side, rear)
    • Collisions with plates, traffic lights, cones and other static obstacles
    • Pedestrian or cyclist being run over

  • The vehicle must travel the route comfortably, in order to allow a pleasant journey for the passenger, avoiding:

    • Abrupt braking
    • Sudden accelerations
    • Sharp curves
    • "Accelerate, brake, accelerate" typical behavior of congested traffic