Conference Proceeding

Why Driving Performance Measures Are Sometimes Not Accurate (and Methods to Check Accuracy)

  • Paul Green (University of Michigan, Ann Arbor)


This paper identifies common sources of inconsistency and error in measurements of driving performance and describes methods to determine the size of these errors. Major sources of inconsistency and error discussed include (1) the lack of zeroing procedures (which affects measurements of steering wheel angle), (2) unknown input and output mapping (which affects measurements of throttle position), (3) the failure to control critical factors such as tire pressure, traffic, and wind (which affects measurements of speed), (4) uncertainty about where the lane boundary actually is (which affect measurements of lane position and counts of lane departures), and (5) the failure to define or use consistent definitions for measures such as headway/gap, time to line crossing, and time to collision. The lack of or inconsistency of definitions can lead to multiple interpretations of what could have been measured, and differences between interpretations are of practical significance. The types and magnitude of these inconsistencies and errors vary with the measurement platform, complicating the comparison of driving studies and interfering with building a body of knowledge of driving. By making the driving research community aware of these problems, they can be identified, assessed, and minimized in the future, and published research can be read with a more critical eye.

How to Cite:

Green, P., (2007) “Why Driving Performance Measures Are Sometimes Not Accurate (and Methods to Check Accuracy)”, Driving Assessment Conference 4(2007), 394-400. doi:

Rights: Copyright © 2007 the author(s)

Download pdf
View PDF



Published on
11 Jul 2007
Peer Reviewed