Skip to content
@SmartDrive-UniPi

SmartDrive-UniPi

Smart Drive - Teleoperated & Autonomous Driving 🏎️📷🎓

Docente referente: Marco Gabiccini, Francesco Frendo

[ITA] 🇮🇹

Il progetto, rivolto agli studenti e alle studentesse del Corso di Laurea Magistrale in Ingegneria dei Veicoli e con il coinvolgimento di studenti tutor della LM in Ingegneria Robotica e dell’Automazione, consiste nello sviluppo di un sistema di guida remota di un veicolo in scala mediante simulatore di guida.

I partecipanti dovranno modificare un veicolo in scala 1:5 radiocomandato dotato di motore elettrico, installandovi un sistema di visione basato su telecamera ad alta definizione e lidar e di un sistema di controllo dell'acceleratore/freno-motore e dello sterzo mediante sistema aperto VESC.

Il sistema dovrà essere teleoperato a distanza tramite la rete 4G LTE o 5.8GHz ed in modalità BVLoS (ossia senza contatto visivo). Questo motiverà la necessità per il veicolo di essere controllato attraverso un simulatore di guida di tipo compatto dotato almeno di volante, pedali e sistema di visione immersivo, da realizzarsi mediante monitor curvi o visore.

Le immagini della telecamera e del lidar verranno inviate al sistema di visione del pilota. Questi, agendo sullo sterzo e sui pedali del simulatore di guida, invierà i suoi comandi al veicolo. La possibilità di aggiungere una scheda tipo NVIDIA AGX Xavier o simili a bordo del veicolo insieme ad un sistema di localizzazione GPS, sistema IMU, ruote foniche ecc. consentirà, altresì, la possibilità di effettuare la guida autonoma del veicolo mediante un sistema di intelligenza artificiale basato su reti neurali profonde (Deep Neural Network) per il riconoscimento dei bordi strada e di una opportuna logica di navigazione. Si prevede anche l’acquisizione di grandezze utili per la stima dello stato di sollecitazione e di danneggiamento di alcuni componenti del veicolo stesso e di grandezze centrali nello studio della dinamica veicolo.

[ENG] 🇬🇧

The project, aimed at students of the Master's Degree Course in Vehicle Engineering and with the involvement of student tutors from the Master's Degree in Robotics and Automation Engineering, consists in the development of a remote driving system for a scale vehicle using a driving simulator.

Participants will have to modify a 1:5 scale radio-controlled vehicle equipped with an electric motor, installing a vision system based on a high-definition camera and lidar, and a control system for the accelerator/brake-motor and steering using an open VESC system.

The system will have to be remotely operated via the 4G LTE or 5.8GHz network and in BVLoS mode (i.e., without visual contact). This will motivate the need for the vehicle to be controlled through a compact driving simulator equipped with at least a steering wheel, pedals, and an immersive vision system, to be implemented using curved monitors or a headset.

The images from the camera and lidar will be sent to the driver's vision system. The driver, acting on the steering wheel and pedals of the driving simulator, will send their commands to the vehicle. The possibility of adding an NVIDIA AGX Xavier or similar board on board the vehicle together with a GPS localization system, IMU system, wheel encoders, etc. will also allow the possibility of performing autonomous driving of the vehicle through an artificial intelligence system based on deep neural networks (Deep Neural Network) for the recognition of road edges and an appropriate navigation logic. It is also expected to acquire quantities useful for estimating the state of stress and damage of some components of the vehicle itself and quantities central to the study of vehicle dynamics.

Pinned Loading

  1. SmartDrive_project SmartDrive_project Public

    Base folder that contain the base documentation and everything needed

    Makefile

  2. navio2_ros navio2_ros Public

    ROS2 Humble packages for Navio2 autopilot shield for Raspberry Pi.

    C++

Repositories

Showing 10 of 15 repositories
  • SmartDrive_project Public

    Base folder that contain the base documentation and everything needed

    SmartDrive-UniPi/SmartDrive_project’s past year of commit activity
    Makefile 0 0 0 0 Updated Oct 23, 2024
  • psd_docker Public

    Docker container configuration with Ubuntu 24.04 Noble, ROS2 Jazzy Jalisco and all the initial dependencies

    SmartDrive-UniPi/psd_docker’s past year of commit activity
    Dockerfile 0 0 0 0 Updated Oct 1, 2024
  • psd_gazebo_sim Public

    Vehicle simulator based on Gazebo Harmonic

    SmartDrive-UniPi/psd_gazebo_sim’s past year of commit activity
    Python 0 0 0 0 Updated Sep 30, 2024
  • psd_slam Public

    Example code stack to locate the vehicle in the world

    SmartDrive-UniPi/psd_slam’s past year of commit activity
    Python 0 0 0 0 Updated Sep 23, 2024
  • psd_DaGGER Public Forked from MatteoMariani99/psd_DaGGER

    psd DaGGER (Dataset Aggregation): algorithm for control based on camera perception and detection of cones or stripes. Control based on a mix of an Imitation Learning alg and an expert driver (in this case a Proportional control)

    SmartDrive-UniPi/psd_DaGGER’s past year of commit activity
    Python 0 1 0 0 Updated Sep 23, 2024
  • gz_ros2_control Public Forked from ros-controls/gz_ros2_control

    Connect the latest version of Gazebo with ros2_control.

    SmartDrive-UniPi/gz_ros2_control’s past year of commit activity
    C++ 0 Apache-2.0 86 0 0 Updated Sep 22, 2024
  • ros_components_description Public Forked from husarion/ros_components_description

    URDF models of sensors and other components offered alongside with Husarion robots

    SmartDrive-UniPi/ros_components_description’s past year of commit activity
    Python 0 6 0 0 Updated Sep 22, 2024
  • psd_perception Public

    Example code stack to emulate camera/lidar perception

    SmartDrive-UniPi/psd_perception’s past year of commit activity
    C++ 0 Apache-2.0 0 0 0 Updated Sep 19, 2024
  • cone_detection Public Forked from MatteoMariani99/cone_detection

    cone detection & segmentation

    SmartDrive-UniPi/cone_detection’s past year of commit activity
    Python 0 1 0 0 Updated Sep 15, 2024
  • gtsam Public Forked from borglab/gtsam

    GTSAM is a library of C++ classes that implement smoothing and mapping (SAM) in robotics and vision, using factor graphs and Bayes networks as the underlying computing paradigm rather than sparse matrices.

    SmartDrive-UniPi/gtsam’s past year of commit activity
    C++ 0 769 0 0 Updated Sep 10, 2024

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…