Vehicle tracking units (VTUs) have become popular in the last decade due to the pervasiveness of their applications, such as fleet management, usage-based insurance and road side assistance. In all such applications it is crucial that the VTUs are continuously working, in order to avoid disruption of the service to the end users. Whenever a VTU stops reporting for a long period of time, e.g. a week, it is thus of extreme importance to be able to understand whether the unit is healthy or in a faulty condition.The present paper proposes a method that automatically detects VTUs that stop reporting under anomalous conditions, for instance due to low data signal coverage, connection loss, power loss, wiring issues or hardware faults; the method separates the aforementioned cases from healthy units that are in a deep sleep mode after the vehicle has been turned off. Our method exploits patterns extracted from the messages sent on the days before the last message, from historical information about the vehicle activity and from contextual information.The discrimination method is a Random Forest classifier, trained on historical VTU message data that include examples of both healthy shut-downs and anomalies, collected by Verizon Connect. A large set of features is built base on such data and a feature importance analysis is then exploited to select the most effective subset of features for the classification model.

Fault detection in non-reporting Vehicle Tracking Units

Salti Samuele;
2019

Abstract

Vehicle tracking units (VTUs) have become popular in the last decade due to the pervasiveness of their applications, such as fleet management, usage-based insurance and road side assistance. In all such applications it is crucial that the VTUs are continuously working, in order to avoid disruption of the service to the end users. Whenever a VTU stops reporting for a long period of time, e.g. a week, it is thus of extreme importance to be able to understand whether the unit is healthy or in a faulty condition.The present paper proposes a method that automatically detects VTUs that stop reporting under anomalous conditions, for instance due to low data signal coverage, connection loss, power loss, wiring issues or hardware faults; the method separates the aforementioned cases from healthy units that are in a deep sleep mode after the vehicle has been turned off. Our method exploits patterns extracted from the messages sent on the days before the last message, from historical information about the vehicle activity and from contextual information.The discrimination method is a Random Forest classifier, trained on historical VTU message data that include examples of both healthy shut-downs and anomalies, collected by Verizon Connect. A large set of features is built base on such data and a feature importance analysis is then exploited to select the most effective subset of features for the classification model.
2019 IEEE Intelligent Transportation Systems Conference, ITSC 2019
2791
2796
Bravi Luca; Simoncini Matteo; Taccari Leonardo; Sarti Leonardo; Caprasecca Stefano; Benericetti Andrea; Lori Alessandro; Salti Samuele; Sambo Francesco
File in questo prodotto:
Eventuali allegati, non sono esposti

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: http://hdl.handle.net/11585/740306
 Attenzione

Attenzione! I dati visualizzati non sono stati sottoposti a validazione da parte dell'ateneo

Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 0
  • ???jsp.display-item.citation.isi??? 0
social impact