Summarizing Normative Driving Behavior From Large-Scale NDS Datasets for Vehicle System Development

2025-07-18

AI Summary

This study addresses the challenge of modeling normative driving behavior using naturalistic driving data. We propose a scalable, cross-cohort behavioral analysis framework leveraging the SHRP 2 dataset—comprising 3,400 drivers and 34 million vehicle miles—integrating onboard sensor, GPS, and forward radar data. Five core metrics are quantified: speed, speeding frequency, lane-keeping stability, following distance, and time headway; these are jointly modeled with road geometry, vehicle type, and driver demographic variables via multivariate statistical analysis and interactive visualization. A novel web-based interactive platform enables dynamic, cross-cohort behavioral comparisons. Critically, we establish the first engineering-oriented normative driving benchmark, revealing that younger drivers exhibit significantly higher speeding rates and shorter time headways. The findings provide empirically grounded insights for ADAS development, autonomous vehicle validation, and road infrastructure optimization.

Technology Category

• Application Domains: Mobility, Driving & Flight
• Planning, Routing, and Scheduling: Model-Based Reasoning
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• User Modeling, Personalization and Recommendation: Metrics for user behavior and evaluating success
• Security and Privacy: Large-scale security measurements
• Responsible Web: Ethical and legal aspects of web-scale data analysis, uses and collection practices

Abstract

This paper presents a methodology to process large-scale naturalistic driving studies (NDS) to describe the driving behavior for five vehicle metrics, including speed, speeding, lane keeping, following distance, and headway, contextualized by roadway characteristics, vehicle classes, and driver demographics. Such descriptions of normative driving behaviors can aid in the development of vehicle safety and intelligent transportation systems. The methodology is demonstrated using data from the Second Strategic Highway Research Program (SHRP 2) NDS, which includes over 34 million miles of driving across more than 3,400 drivers. Summaries of each driving metric were generated using vehicle, GPS, and forward radar data. Additionally, interactive online analytics tools were developed to visualize and compare driving behavior across groups through dynamic data selection and grouping. For example, among drivers on 65-mph roads for the SHRP 2 NDS, females aged 16-19 exceeded the speed limit by 7.5 to 15 mph slightly more often than their male counterparts, and younger drivers maintained headways under 1.5 seconds more frequently than older drivers. This work supports better vehicle systems and safer infrastructure by quantifying normative driving behaviors and offers a methodology for analyzing NDS datasets for cross group comparisons.

Problem

• Summarize normative driving behavior from large-scale NDS datasets
• Develop methodology for analyzing driving metrics across diverse groups
• Support vehicle safety and intelligent transportation systems development

Innovation

• Process large-scale NDS data for driving behavior
• Use vehicle, GPS, radar data for metric summaries
• Develop interactive tools for behavior visualization