In-vehicle data recorders for monitoring and feedback on drivers’ behavior

https://doi.org/10.1016/j.trc.2008.01.001Get rights and content

Abstract

This paper describes the potential of in-vehicle data recorder (IVDR) systems to be used in various commercial and research applications as tools to monitor and provide feedback to drivers on their on-road behavior. The implementation of IVDR is demonstrated using the example of the DriveDiagnostics system. This system can identify various maneuver types that occur in the raw measurements, and use this information to calculate risk indices that indicate on the overall trip safety. Drivers receive feedback through various summary reports, real-time text messages or an in-vehicle display unit. Validation tests with the system demonstrate promising potential as a measurement tool to evaluate driving behavior. Reductions in crash rates and the risk indices are observed in the short-term.

Introduction

The human and cost implications of car crashes are staggering. For example, Blincoe et al. (2002) estimated the total direct annual cost of car crashes in the US in 2000 at $230.6 Billion, and the total cost to society at $493.3 Billion. The direct cost of a car crash was estimated at $14,000, out of which $3600 is the cost of damage to vehicles and other property. For company vehicles, various studies (Fidderman, 1993, Murray et al., 1996, Lynn and Lockwood, 1999) estimated that 20–65% will be involved in car crashes each year. Lynn and Lockwood (1999) also found that even after controlling for the larger distances they drive, company car drivers are 50% more likely to be involved in car crashes compared to other drivers. Thus, it is clear that the potential benefits of tools and methods that can reduce the rate of involvement in car crashes are huge, and that this potential is in particular substantial for vehicle fleets.

Driver behavior and errors are a cause in the overwhelming majority of car crashes (Evans, 2004). Understanding and influencing drivers’ behavior is therefore essential in order to improve road safety. An important obstacle to better understanding of drivers’ behavior is data availability. Most tools used to define drivers’ skills and styles are based on self-reported scales, which could be biased and general (see Podsakoff and Organ, 1986 for a discussion). Otherwise, evaluations of driving behavior may only be performed on a limited scope and scale, for example in experiments involving driving simulators (Comte, 2000).

Recently, In-Vehicle data recorders (IVDR) have emerged as new tools to collect data on driving behavior and to provide feedback to drivers continuously, in much more detail and with large scale implementations. IVDR are on-board devices that collect and record information on the movement, control and performance of the vehicle (NHTSA, 2001, Chidester et al., 2001, Correia et al., 2001). The technology was first used in event data recorders (EDR), which store information on the states of the vehicle’s systems for a short time (about 30 s) before, during and after crash events. This information is used to evaluate and improve safety equipment and to investigate crash causes and allocate fault (see NHTSA 2005 for a thorough review). Limited empirical evidence suggests that the installation of EDR may affect drivers’ behavior. Lehmann, 1996, Lehmann and Cheale, 1998 report reductions of 15–30% in crash rates, and even more significant reductions in the related costs, in several vehicle fleets that installed EDR. In all cases, drivers did not receive any feedback, but were aware of the EDR presence. Wouters and Bos (2000) report an overall reduction of 20% in crash rates in several truck, bus and taxi fleets that installed EDR. Drivers received feedback from the EDR after a crash has occurred. In contrast, Heinzmann and Schade (2003) found that EDR did not have any significant impact on the behavior or on crash rates of young males.

Several studies focused on development and evaluation of systems that monitor drivers’ behavior continuously and not only in crash events. NHTSA (Neale et al., 2002, Dingus et al., 2006) conducted an ambitious study in which 100 vehicles were instrumented with IVDR that continuously measure the vehicle’s position, speed and acceleration using GPS and accelerometers as well as video cameras showing the inside and outside of the vehicle, radar sensors and lane trackers. The system also extracts information from the vehicle’s on-board diagnostics system. The experiment was conducted over a 13 months and yielded a data set with over 43,000 h and 2 million miles driven. This data has great potential for traffic safety research. However, instrumentation at this level is costly and so unlikely to be deployed in large scale in the near future.

Ogle (2005) used an IVDR system to collect data from 172 vehicles. The system incorporates a GPS receiver and is able to connect to the vehicle’s on-board diagnostics system. It can collect the time and durations of trips, distance traveled, second-by-second position, speed and acceleration as well as various engine parameters and information on seat belt usage, emissions and brake and throttle positions. Trip level information was stored in the unit and transmitted to the application server once a week through a wireless network. The availability of location data allowed identification of geographic information, such as types of roads traveled and posted speed limits. The system does not provide any feedback to drivers. While the author points out a wide range of studies the data may be used for, the analysis she presents was limited to a model that predicts the maximum speeding level above the speed limit in individual trips as a function of the driver’s socio-demographic characteristics, the type of road facility and trip characteristics (e.g. time of day, level of congestion). Within this model, there was no significant relation between speeding behavior and past crash records.

Commercial applications of IVDR have also emerged in recent years. The TripSense program (TripSense, 2007) uses IVDR data to determine insurance rates for participating vehicles. Installation of the system entitles drivers for a discount on their insurance premium. The discount level is determined by the vehicle utilization pattern (i.e. the hours and distance driven and their distribution over the day and the week) and the speed profile. Speeds are collected at 10 s resolution, but location information is not collected, and so speed limits are not known. The speed factor is therefore expressed by the fraction of time speed is over 75 mph. The system also collects, but does not use, information on the occurrence of sudden starts and stops. Results demonstrating the impact on drivers’ behavior are not presented. ECMT (2006) report on the SAGA system developed in Iceland. This IVDR collects information on the vehicle utilization, speed and location using GPS. The location data allows comparison of travel speeds with posted speed limits. Weekly summary reports are sent to users by email. Installation of this system Iceland Post vehicles resulted in a 43% reduction in crash rates over a period of six months. The number of vehicles instrumented and the details of the feedback drivers received were not reported.

In this paper, we describe the details of a specific IVDR system and report several results that demonstrate its potential to measure and impact drivers’ behavior: We find that risk indices that the system calculates are significantly correlated with past crash involvement. This result suggests that these indices may be used as indicators to the risk of crash involvement. We also find significant reductions in crash rates after the installation of the IVDR, and use the risk indices to study the temporal changes in the impact of the IVDR installation and feedback. Finally, we discuss potential applications of the technology, both commercial and academic.

Section snippets

The Drivediagnostics system

The overall framework of the system, which is shown in Fig. 1, comprises of four different tasks: measurement, detection, analysis, and feedback.

Experiment

The results reported here are based on installations in 191 vehicles in a single company. The vehicles are all compact pickup trucks that are used only on the job, by technical employees that use them to travel between service locations. Each vehicle is assigned to a single driver, and vice versa. Although they drive significant mileage, the drivers (189 males, 2 females) are not professional drivers and are not employed as drivers. The implementation of IVDR systems in all these vehicles

Connection between IVDR risk indices and crash involvement

The risk indices computed by the IVDR may be useful, if they are correlated with the actual risk of involvement in car crashes. However, the “true” risk cannot be directly measured and so we use past crash records as indicators and test the hypothesis that they are correlated with IVDR risk indices. Crash records were collected for all 191 drivers whose vehicles were instrumented with the IVDR. These drivers are between 25 and 68 years old, with an average age of 41 years. The IVDR risk indices

Conclusion

The results presented above demonstrate the potential usefulness of IVDR systems, which may be important both in commercial applications and in research. In this section we list several current and potential applications of the technology. In these applications, the IVDR has two roles: Firstly, it is a tool to objectively measure and evaluate driving behaviors. Secondly, it can be used to impact drivers’ behaviors through monitoring and provision of feedback.

IVDR have important advantages in

Acknowledgement

The authors wish to thank three anonymous referees for their many helpful comments and suggestions.

References (23)

  • M. Roetting et al.

    When technology tells you how you drive – truck drivers’ attitudes towards feedback by technology

    Transportation Research Part F

    (2003)
  • P.I.J. Wouters et al.

    Traffic accident reduction by monitoring driver behaviour with in-car data recorders

    Accident Analysis and Prevention

    (2000)
  • Blincoe, L., Seay, A., Zaloshnja, E., Miller, T., Romano, E., Lutcher, S., Spicer, R., 2002. Economic cost of motor...
  • Chidester, A.C.D., Hinch, J., Roston, T.A., 2001. Real world experiences with event data recorders. In: Proceedings of...
  • L.S. Comte

    new systems: new behaviour?

    Transportation Research Part F

    (2000)
  • Correia, J.T., Iliadis, K.A., McCarron, E.S., Smolej, M.A., 2001. Utilizing data from automotive event data recorders....
  • Dingus, T.A., Klauer, S.G., Neale, V.L., Petersen, A., Lee, S.E., Sudweeks, J., Perez, M.A., Hankey, J., Ramsey, D.,...
  • ECMT

    Speed management

    (2006)
  • L. Evans

    Traffic Safety

    (2004)
  • H. Fidderman

    Company car driver behaviour

    Health and Safety Information Bulletin

    (1993)
  • Heinzmann, H.J., Schade, K.B., 2003. Moderne Verkehrssicherheits Technologie Fahrdatenspeicher und Junge Fahrer....
  • Cited by (0)

    View full text