This study evaluates the accuracy of 41 Event Data Recorders
(EDR) extracted from model year 2012 General Motors, Chrysler,
Ford, Honda, Mazda, Toyota, and Volvo vehicles subjected to New Car
Assessment Program 56 kph full-frontal barrier crash tests. The
approach was to evaluate (1) the vehicle longitudinal change in
velocity or delta-V (ΔV) as measured by EDRs in comparison with the
high-precision accelerometers mounted onboard test vehicles and (2)
the accuracy of pre-crash speed, seatbelt buckle status, and
frontal airbag deployment status.
On average the absolute error for pre-crash speed between the
EDR and reference instrumentation was only 0.58 kph, or 1.0% of the
nominal impact speed. In all cases in which the EDRs recorded the
seatbelt buckle status of the driver or right front passenger, the
modules correctly reported that the occupants were buckled. EDRs
reported airbag deployment correctly in all of the tests. The
average absolute error was 4.20 kph (6.6%) for final longitudinalΔV
and 4.32 kph (6.6%) for maximum longitudinal ΔV. Our results show
that EDRs underreport the reference instrumentation ΔV in the vast
majority of cases.