The need to improve road safety performance for all road users is clear, particularly for vulnerable road users (such as pedestrians and cyclists), and users of micro-mobility services (such as e-scooters). The optimization of investment by local and state agencies to maximize lives saved and injuries reduced takes on even greater importance when financial resources are constrained. Unlocking the broader sustainable benefits that come from active transportation modes also requires an understanding of the safety performance of infrastructure. The absence of low-cost data, safety performance metrics, and prioritized investment options make it difficult for agencies to understand the business case for safer roads and to measure progress.
This research will investigate the use of artificial intelligence (AI), machine learning (ML) and Big Data (BD) to provide the information needed to power key data-driven, public and proprietary safety analysis tools as well as predictive and other systemic safety tools. The availability of large-scale and consistently collected data across the entire road network will improve the visibility of existing network conditions with a focus on road and exposure features influencing the safety of all road users. This low-cost and consistent data can then inform and accelerate the investments needed to support safe system outcomes with a particular focus on modal priority and the needs of pedestrians, cyclists, and new-mobility users. Even with the introduction of connected and automated vehicles (CAV), investments to ensure an efficient and optimal interaction between all road users will continue to need to prioritize vulnerable road users.
The research will build on the AI innovations under development globally for Road Assessment Programs (RAP) in other countries. AI-RAP captures the advances in AI, ML, vision systems (street and sky), light detection and ranging (LiDAR), telematics, and other data sources to deliver critical information on road safety, crash performance, investment prioritization, and RAP’s Star Rating of roads for pedestrians, cyclists, motorcyclists, and vehicle occupants. The accelerated and intelligent coding of these attributes can provide significant savings and deliver the scale and frequency of data collection and analysis to support comprehensive performance tracking over time.
The objective of this research is to advance the use of AI and ML in analyzing BD and unconventional data and assessing their effectiveness to support safe system and modal priority decision-making as well as performance tracking. The resultant algorithms are expected to improve and optimize analyses using existing data and data-driven safety analysis tools developed based on conventional statistical modeling (see, for example, NCHRP Research Report 955: Guide for Quantitative Approaches to Systemic Safety Analysis).
Note: Assessing the effectiveness of BD and unconventional data might include, for example, determining biases in the data or identifying data that do not represent an entire population.
The research will also (a) identify potential data sources, (b) identify or develop the requisite data preparation and extraction algorithms for use in safety analysis, and (c) document each source’s coverage, frequency of collection, granularity, accessibility to practitioners, and cost. These sources shall include but not be limited to video data, telematics, LiDAR, satellite, aerial imagery, weather, land use, location-based services data, crowd-sourced data, and demographic and census data. This data will allow the potential for lower-cost and more frequent generation of, among others: key fatality and injury prediction risk maps; road feature mapping; star ratings and other safety analyses for pedestrians, cyclists, motorcyclists, micro-mobility services, and vehicle occupants; identification of data for safety analyses and associated tools; and the development of safety plans that can be used for funding submissions and in prioritizing investments across the local and state road networks.
Finally, this research will develop guidance for managing data using a format that can be accessed by various tools. This guidance should be tested through pilot projects to allow for appropriate adjustment and greater understanding. The development of guidance will enhance implementation and provide necessary information on the use of this data in safety systems and in determining modal priority needs. Results of this research could be included in national-level resources such as the AASHTO Highway Safety Manual and other tools that support data-driven safety analysis.
STATUS: Research in progress.