Search
Close this search box.

RDI PROJECTS

CL-VCON – Clustering Connected Vehicles

The project is under development since 2022
POC: innovacion@isdefe.es

Description

Winner of the 2021 Isdefe’s IDEAS challenge (7th edition)

Internal R&D project

Isdefe’s internal R&D project Clustering of Connected Vehicles (CL-VCON) aims to obtain useful information about a fleet of vehicles by applying clustering techniques (unsupervised machine learning algorithms).

Data from the vehicles are collected by means of sensors installed in the vehicles. The information from the sensors is stored for further analysis using clustering algorithms.

The project will result in vehicle information related to the following concepts:

  • Detect anomalous or defective vehicles
  • Anticipate breakdowns by benchmarking vehicles with similar behaviour.
  • Detect vehicle misuse and driving deficiencies.
  • Environmental recommendations for driving (consumption, gas emissions).
  • Graphical display of connected vehicles according to their similarities.

The advantage of applying such algorithms is that less time is needed to collect data from the vehicles and it is not necessary to have a vehicle maintenance history in order to achieve useful results for predictive vehicle maintenance.

The project is being developed in the UME with a fleet of 25 Renault K-430 vehicles. The contract is currently being awarded to the company that will carry out the instrumentation of the vehicles.

In this project, data is downloaded from the vehicles manually, but everything will be prepared to allow for unattended wireless downloading of data at a later stage.

It is intended that this project will be the basis for further development of AI-based projects in the EMU. Once the vehicles have been sensorised, this data can be used for further predictive analysis.

In addition to the initial proposal, some further technological innovations will be incorporated into the project:

  • Unsupervised anomaly detection algorithms.
  • Joint use of Excel (Visual basic) and Python.
  • Generation of synthetic data for the development of AI models, in the event that we do not have sufficient data for a given environment.

The project is being developed by a team coordinated by Gonzalo Santamaria Freire and Javier Castrillo Mazagatos of Isdefe’s Defence and Security Technologies Division.

For further information, please contact innovacion@isdefe.es

If you liked this content, you can share it in your networks