Tesi etd-01172020-152549
Link copiato negli appunti
Tipo di tesi
Master di Secondo Livello
Autore
CUPELLINI, FEDERICO
URN
etd-01172020-152549
Titolo
Retrospective analysis-method for estimating the effect of Advanced
Driver Assistance Systems on the costs of car crashes
Struttura
Istituto di Management
Corso di studi
Corsi Alta Formazione - MASTER IN MANAGEMENT OF INNOVATION - MAINS
Commissione
relatore Prof. RAPACCINI, MARIO
Tutor Ing. IANNELLO, CALOGERO
Tutor Ing. IANNELLO, CALOGERO
Parole chiave
- dynamic capabilities
- road safety
- sicurezza stradale
Data inizio appello
07/02/2020;
Disponibilità
completa
Riassunto analitico
Advanced driver assistance systems (ADAS) are a type of technologies fitted on vehicles designed to improve
safety. ADAS help drivers through warnings or autonomous intervention to avoid dangerous situations or
mitigate the effects of a crash. Laboratory tests suggest that ADAS can provide a marked reduction in crash
rates, however real-world evidence on ADAS effectiveness is limited. This study investigates a strategy to
analyze the impact of ADAS technologies presence on the likelihood of crashes and the monetary impact for
the fleet of the largest long-term rental provider in Italy. Vehicle ADAS presence, and characteristics for cars
registered since 2016, was identified through quotation-coded options data (n=1,987). ADAS technologies
considered include lane keeping assist, parking assist and emergency braking. Arval accident management
data (from January 2016 to August 2019) were merged with vehicle data by VIN, identifying vehicles in-
volved in accidents (n=483). Data included accident insurance class and total cost of repairs. Using Cox
proportional hazards regression modeling (following (Chang and Jovanis 1990)), we calculate the adjusted
hazard ratio for accidents among Arval vehicles equipped versus not equipped with ADAS. The percentage
reduction in crashes related to ADAS presence is interpreted as one minus the hazard ratio. Moreover a
multivariate analysis is performed, creating a model to predict accidents occurence and crash costs. Conclu-
sion: This study provides a methodology for the analysis of ADAS technologies effectiveness in the fleet of
a major Italian long term vehicle rental provider.
safety. ADAS help drivers through warnings or autonomous intervention to avoid dangerous situations or
mitigate the effects of a crash. Laboratory tests suggest that ADAS can provide a marked reduction in crash
rates, however real-world evidence on ADAS effectiveness is limited. This study investigates a strategy to
analyze the impact of ADAS technologies presence on the likelihood of crashes and the monetary impact for
the fleet of the largest long-term rental provider in Italy. Vehicle ADAS presence, and characteristics for cars
registered since 2016, was identified through quotation-coded options data (n=1,987). ADAS technologies
considered include lane keeping assist, parking assist and emergency braking. Arval accident management
data (from January 2016 to August 2019) were merged with vehicle data by VIN, identifying vehicles in-
volved in accidents (n=483). Data included accident insurance class and total cost of repairs. Using Cox
proportional hazards regression modeling (following (Chang and Jovanis 1990)), we calculate the adjusted
hazard ratio for accidents among Arval vehicles equipped versus not equipped with ADAS. The percentage
reduction in crashes related to ADAS presence is interpreted as one minus the hazard ratio. Moreover a
multivariate analysis is performed, creating a model to predict accidents occurence and crash costs. Conclu-
sion: This study provides a methodology for the analysis of ADAS technologies effectiveness in the fleet of
a major Italian long term vehicle rental provider.
File
| Nome file | Dimensione |
|---|---|
| PW_Cupel...Mains.pdf | 2.78 Mb |
Contatta l'autore |
|