Coherence assessment of accident database kinematic data

Area of Science:

• Road safety research
• Accident reconstruction
• Data analysis

Background:

• Accident analysis relies on kinematic data from databases to assess vehicle safety and injury risk.
• Ensuring data accuracy and representativeness is crucial for reliable analysis.
• Current methods rarely check kinematic data coherence during the collection phase.

Purpose of the Study:

• To present a procedure for verifying the internal coherence of kinematic data in accident databases.
• To identify kinematic parameters that are inconsistent due to inappropriate accident reconstruction models.
• To improve the quality and reliability of accident data for safety research.

Main Methods:

• Developed checks based on fundamental physical laws: momentum conservation, velocity triangle compatibility, and energy conservation.
• Applied the procedure to vehicle-to-vehicle collision data in two real-world databases.
• Methodology is adaptable to other parameters and databases.

Main Results:

• The procedure successfully identified a non-negligible number of incongruent kinematic data points in real databases.
• Inconsistent data can significantly impact direct analyses (e.g., injury risk curves) and secondary analyses.
• The checks are vital for confirming data congruence during both data collection and analysis phases.

Conclusions:

• Implementing kinematic data coherence checks is essential for enhancing the quality of accident investigations.
• Accurate data improves the reliability of injury risk assessments and vehicle safety system evaluations.
• The proposed methodology offers a robust approach to data validation in accident research.