Amplifying Traditional Safety Assessments: From Reactive to Predictive

Safety is one of TxDOT’s priorities, and to support those efforts, ConnectSmart has introduced numerous features, including Low Clearance Structure Warnings, Flood Alerts, and integration with Cellular Vehicle-to-Everything (C-V2X) ―a cutting-edge communications technology that enables vehicles to interact with their surroundings, including other vehicles, pedestrians, cyclists, and infrastructure―to enhance safety. In addition to system integrations, user data acquired through the ConnectSmart app can further support TxDOT’s safety goals and objectives. To that extent, ConnectSmart introduced a methodology to calculate indicators derived from telematics so that traffic engineers can identify locations with elevated systemic safety risk.

Historically, roadway safety analysis has relied heavily on a foundation of lagging indicators (such as crash and fatality rates) to understand historical performance across network components. While these indicators are foundational to identifying safety trends, they represent events that have already occurred. By integrating critical leading indicators and utilizing vehicle trajectory data, ConnectSmart’s methodology transitions safety analysis from a purely reactive model to a highly proactive one.

The methodology enhances traditional “event-based” metrics with "style-based" metrics—smoothness and consistency- to capture the behavioral baseline of the motorist and the systemic turbulence of the roadway. Individual trips are aggregated into a standardized roadway segment score to isolate systemic operational turbulence. A consistently low score on a specific segment across a high volume of drivers indicates an underlying infrastructure deficiency—such as suboptimal signal timing or geometric design ambiguities—rather than isolated driver error. This predictive score is designed to be utilized in conjunction with traditional metrics, including historical crash data, geometric alignment, access-point density, and signal spacing, to prioritize corridors for proactive roadway safety assessments before crash frequencies escalate.

Individual Score: Driver Behavior

The Individual Driver Score (IDS) is calculated on a 100-point scale and comprises three core components:

●      Event-Based Metrics: Measures the frequency and intensity of critical maneuvers, including speeding, hard braking, and rapid acceleration over time.  

●      Smoothness: Quantifies speed variability, serving as a proxy for following distance. Safe driving profiles maintain a relatively constant velocity, whereas high speed variability serves as a mathematical signature for latent risks like tailgating and driver distraction.

●      Consistency: Evaluates lateral vehicle stability by analyzing heading deviation and steering reversal rates. It serves as the primary mechanism for detecting micro-behaviors associated with lane weaving, driver fatigue, or impairment.

Scaling Up: The Roadway Composite Score (RCS)

While individual driver scores assess personal operating habits, scaling these data points to a defined roadway segment yields the Roadway Composite Score (RCS).  

If a specific segment yields a consistently low score across a statistically significant sample of drivers, it points to an infrastructure-level issue—such as poor signal progression or restricted sight distance—rather than localized driver error. This aggregated data serves as a strategic screening tool to prioritize segments for formal safety audits and countermeasure selection.

Real-World Application: Houston's High-Risk Corridors

Throughout March 2026, this methodology was deployed across the Houston market on three select roadway segments, including a freeway frontage road, a farm-to-market road principal arterial, and a downtown collector with an active work zone.

Freeway Frontage Road

Categorized as a "Critical" segment due to the crash volumes within the Houston District, this corridor recorded an Event Score of 46.07 and an overall RCS of 74.24. Notably, the Smoothness and Consistency metrics were 90.21 and 96.97, respectively, which did not deviate significantly from regional baselines, indicating that safety risks along this corridor are heavily driven by discrete, high-intensity conflict events rather than continuous systemic turbulence.

Farm-to-Market Road/ Principal Arterial

As the highest-ranked off-system (non-interstate) corridor for pedestrian crashes in the region (referencing the TxDOT Houston District Pedestrian Safety Profile and Vision Zero Houston’s High Injury Network), this segment recorded an Event Score of 65.66 and an overall RCS of 82.60. Similar to the I-45 corridor, Smoothness and Consistency scores were 93.44 and 94.54, respectively, confirming that safety risks are primarily event-driven.

Downtown Collector/ Active Construction Zone

An evaluation was conducted on the Main Street Promenade downtown project, which commenced in mid-2025 and featured partial road closures and modified vehicle access. In the segments under construction, the overall RCS dropped from a pre-construction baseline of 84.80 to 80.98 during the active construction phase. However, data indicated that while both the Event (from 72.80 to 69.97) and Smoothness (from 97.01 to 84.34) scores dropped, the Consistency score increased (from 86.89 to 93.88) within the construction zone, a phenomenon likely attributable to naturally less weaving and higher driver alert dictated by construction-related congestion and messaging.

The Path Forward for Safety

By quantifying both motorist behavioral profiles and systemic roadway turbulence, this framework successfully distinguishes between localized driver error (individual score degradation) and systemic infrastructure deficiencies (global score degradation across all users). Integrating this proactive methodology into traditional safety workflows enhances location prioritization, allowing transportation agencies to allocate life-saving capital resources and systemic countermeasures where they will yield the highest reduction in fatal and injury crashes.