Distributed machine learning on vehicles
keywords MACHINE LEARNING, VEHICULAR NETWORKS
Reference persons CLAUDIO ETTORE CASETTI, PAOLO GIACCONE
Research Groups Telecommunication Networks Group
Thesis type DEVELOPMENT, SIMULATION
Description Given the plethora of sensors with which vehicles are equipped, today’s automated vehicles already generate large amounts of data that enable data-driven solutions for vehicle control, safety and comfort, as well as to effectively implement convenience applications.
The goal of this thesis work is to develop a custom framework on the CARLA simulator (https://carla.org/) leveraging SUMO traces of vehicles, and other transportation means. The framework would exploit the use of different technologies (i.e., camera, lidar, radar, IMU, etc.) and distributed machine learning techniques (e.g., federated learning) in order to develop object detection applications focused on crowdsensing tasks directly from the vehicles video cameras.
Required skills Background on machine learning
Excellent programming in Python
Deadline 20/12/2023 PROPONI LA TUA CANDIDATURA