Sviluppo di tecniche di modellazione, ottimizzazione e predizione avanzate, in ambito veicolare, sulla base di dati di telematica e da sistemi IoT

PhD in Pure and Applied Mathematics


Francesco Vaccarino –
Riccardo Loti –

PhD Student: Silvia Buccafusco

Context of the research activity

Optimizing the usage of a fleet equipment is one of the key aspects that can be improved by the means of a telematic system. The telematics system of Tierra records several data of each monitored vehicle, including each on‐off cycle, the intensity of the work done, engine statistics and other mechanical parameters of the vehicle.
The scientific activity aims at studying the usage of the monitored equipment, the recurring patterns, the seasonality, and spatial tendencies by means of statistical/Machine Learning models. The study is supported by both information from Tierra’s DB (e.g. fleet) and from external data sources. Specifically, in this project we
consider different features in input related to i) the vehicle usage pattern (fuel consumption, travelled distance, location, etc.) and ii) external features (weather, sunshine duration, calendar schedule, holidays, etc.), iii) features from other vehicles working in the same site (usage of similar vehicles, usage of nearby vehicles, etc.).
The Company Tierra has planned for the winner of this position a collaboration within a contract of high apprenticeship according to the Italian Legislative Decree 81/2015, art. 45.


The aim of this project is twofold, namely:

  • Sample and compress fine‐grained vehicle data to optimize IoT data transmission: CAN bus data and miscellaneous vehicle data are transmitted through telematic systems to enable vehicle‐related data monitoring and analysis. The granularity of the transmitted data strongly influences the quality of the performed analyses. To explore the actual potential of Machine Learning and Big Data Analytics
    solutions, Tierra is now interested in performing a drill‐down in order to understand the added value of fine‐grained data.
    To support such analyses, it is crucial to optimize bandwidth usage in CAN bus data transmission, e.g., avoid sending duplicated information, recognize and compress periodical signals, and tailor the sampling rate to the actual needs and contexts. The scientific activity addresses the use of statistical and Machine Learning‐based approaches to time series data analyses to perform data correlation analysis and propose effective solutions for CAN bus data compression, as well as identifying statically and dynamically the optimal sampling methods based on the data source.
  • Explore the added value of fine‐grained data to support intelligent vehicle state definition: According to type, model, and context of usage, industrial vehicles in the fleet are characterized by different usage patterns. Fleet managers are very interested in characterizing the current usage of each vehicle by evaluating the intensity of the work done through its transmitted CAN bus signals. However, the currently available state variables are not based on an in‐depth CAN bus data exploration. The available coarse data only allows to average the behavior of the vehicle over many minutes. Now, the fine‐grained CAN bus data analysis is possible to carry out
    and will be functional to the definition of new, smarter state variables which reflect the detailed vehicle usage under multiple aspects. The scientific activity complements the analyses of fine‐grained data with the definition and analysis of new descriptors of vehicle usage levels, also working toward the integration of diagnostic DM1 information in the mix.

This project will be jointly developed by Tierra S.p.a. and the interdepartmental center SmartData@PoliTo of Politecnico di Torino.

Skills and competencies for the development of the activity

The candidate should have solid competences in computational mathematics. In particular: optimization, mathematical foundation of machine learning, statistics, linear algebra, topology. In computer science: databases, neural networks, programming, big data handling.
Programming: C++, matlab, python, R, tensor flow.
Team working capability is required as well.
The candidate shall be less than 30 years old at the moment of the hiring from the company.

Further information about the PhD program at Politecnico can be found here

Back to the list of PhD positions