PRISMA: A novel approach for deriving probabilistic surrogate safety measures for risk evaluation

Area of Science:

• Road safety engineering
• Traffic conflict analysis
• Predictive modeling in transportation

Background:

• Surrogate Safety Measures (SSMs) assess road safety risks in traffic conflicts.
• Current SSMs often rely on trajectory assumptions, limiting their real-world applicability.
• A need exists for adaptable SSMs that provide real-time risk estimations.

Purpose of the Study:

• Introduce the Probabilistic RISk Measure derivAtion (PRISMA) method for deriving novel SSMs.
• Enable real-time calculation of crash probabilities using a data-driven approach.
• Develop SSMs adaptable to various traffic scenarios without restrictive assumptions.

Main Methods:

• Utilized a data-driven approach to predict future traffic participant trajectories.
• Employed simulations to determine the probability of specific events (e.g., crashes).
• Integrated a regression model with simulation results for real-time risk estimation.

Main Results:

• Derived an SSM for evaluating risk in longitudinal traffic interactions.
• Developed a benchmarking method to assess SSM performance against expected risk trends.
• Demonstrated that the derived SSM aligns with expected risk trends.

Conclusions:

• The PRISMA method offers a flexible, data-driven approach to deriving SSMs for real-time risk assessment.
• The derived SSM shows promise for improving road safety analysis.
• Future research will extend PRISMA to other conflict types, including lateral interactions and vulnerable road users.