3.3 Traffic Safety and Collision Analysis

3.3 Traffic Safety and Collision Analysis

Traffic safety analysis is a foundational component of transportation engineering, systematically addressing the intersection of human behavior, vehicle performance, and roadway design to reduce the frequency and severity of crashes. The NCEES PE Civil Transportation exam heavily tests the ability to evaluate crash data, identify hazardous locations, and apply the Highway Safety Manual (HSM) methodologies. This section provides an in-depth exploration of crash analysis metrics, the safety management process, human factors in highway safety, and Road Safety Audits (RSAs).

Equivalent Property Damage Only (EPDO) Crash Weighting

Raw crash counts rarely provide a complete picture of a site's safety performance because they treat all crashes—from minor fender-benders to fatal collisions—equally. To address this, the Equivalent Property Damage Only (EPDO) method assigns weights to crashes based on their severity, typically scaling them relative to a property-damage-only (PDO) crash.

The severity of a crash is universally categorized using the KABCO scale:

• K (Fatal): One or more persons died within 30 days of the crash.
• A (Incapacitating Injury): Severe injuries preventing the victim from continuing normal activities (e.g., broken limbs, severe lacerations).
• B (Non-incapacitating Injury): Evident injuries that are not disabling (e.g., bruising, minor cuts).
• C (Possible Injury): No visible injury, but complaint of pain or momentary unconsciousness.
• O (PDO - Property Damage Only): No injuries; only vehicle or property damage.

The EPDO rate allows agencies to calculate a composite score for a location. Let $W_K$, $W_A$, $W_B$, $W_C$, and $W_O$ represent the weight factors for each severity level (where $W_O = 1.0$). The EPDO score for a site is calculated as: $EPDO = W_K(N_K) + W_A(N_A) + W_B(N_B) + W_C(N_C) + W_O(N_O)$ Where $N_i$ is the number of crashes of severity $i$. For example, if a fatality is weighted at 150 times a PDO crash, $W_K = 150$. By normalizing crash data by volume (e.g., crashes per million entering vehicles for intersections, or per 100 million vehicle miles traveled for segments), engineers can compare the EPDO rates of different sites to prioritize funding for high-severity locations.

The HSM 6-Step Safety Management Process

The Highway Safety Manual (HSM) outlines a rigorous, data-driven approach to safety management. The Network Screening and Safety Management process consists of six distinct steps:

1. Network Screening: This initial step involves reviewing the entire transportation network to identify and rank sites that are expected to benefit the most from safety improvements. Methodologies range from simple crash frequency to advanced Empirical Bayes (EB) methods that account for regression-to-the-mean.
2. Diagnosis: Once high-priority sites are identified, engineers conduct a detailed review of crash data (using collision diagrams and summary tables) and site conditions to understand the underlying patterns and contributing factors (e.g., a high proportion of rear-end crashes at a signalized intersection).
3. Select Countermeasures: Based on the diagnosis, engineers identify potential engineering solutions. If a site has a high rate of nighttime run-off-road crashes, countermeasures might include installing raised pavement markers, improving lighting, or adding edge-line rumble strips.
4. Economic Appraisal: This step compares the expected benefits of the countermeasures (quantified as the monetary value of crashes prevented) against the costs of implementation and maintenance. Common metrics include the Benefit-Cost Ratio (BCR) and Net Present Value (NPV). A BCR $> 1.0$ indicates an economically viable project.
5. Prioritize Projects: Because agencies have limited budgets, projects are ranked based on their economic appraisal and other strategic goals. This ensures the maximum network-wide safety return on investment.
6. Safety Effectiveness Evaluation: After a countermeasure has been installed for a sufficient period (usually 3-5 years), engineers evaluate its actual performance to determine if the expected crash reductions were achieved. This feedback loop improves future decision-making.

Human Factors in Highway Safety

The driver is the most complex and variable component of the highway system. Understanding human factors is essential for designing roadways that accommodate natural human limitations.

Perception-Reaction Time (PRT): PRT is the time required for a driver to perceive a hazard, recognize the need to act, decide on an action, and initiate that action (e.g., moving the foot to the brake pedal). AASHTO typically recommends a design PRT of 2.5 seconds for standard stopping sight distance (SSD) calculations, which accommodates approximately 90% of all drivers under typical rural conditions. For complex urban environments or unexpected hazards, PRT can be significantly longer (up to 3.0 or 4.0 seconds). The distance traveled during this time is calculated as: $d_r = 1.47 v t$ Where $d_r$ is the reaction distance (ft), $v$ is the vehicle speed (mph), and $t$ is the PRT (sec).

Driver Expectation: Drivers rely heavily on expectations built from past experiences. "A priori" expectations are long-term (e.g., freeway exits are typically on the right). "Ad hoc" expectations are formed immediately by the current driving environment (e.g., a series of gentle curves leads a driver to expect another gentle curve). When a design violates driver expectations—such as a sharp, hidden curve immediately following a long tangent, or a left-hand freeway exit—reaction times increase, and the likelihood of driver error skyrockets.

Visibility and Information Processing: A driver’s visual field narrows as speed increases. Furthermore, humans can only process a limited amount of information simultaneously. Safety requires clear, unambiguous guidance. Signage should be spaced appropriately to prevent information overload. "Positive guidance" involves placing the right information, in the right format, at the right location, giving the driver ample time to react.

Aging Drivers: As the driving population ages, designers must account for physiological changes. Older drivers generally experience degraded visual acuity (especially at night or in glare), slower cognitive processing speeds, restricted neck mobility, and longer PRT. Accommodations for older drivers include larger lettering on signs, better retroreflectivity, improved intersection lighting, and skewed intersections being realigned to 90 degrees to reduce the need for extreme head-turning.

Road Safety Audits (RSAs)

A Road Safety Audit (RSA) is a formal, independent, and comprehensive safety performance examination of an existing or future road or intersection by an independent, multidisciplinary team. Unlike traditional safety reviews which merely check for compliance with design standards, RSAs proactively identify potential safety issues for all road users under all conditions.

RSAs can be conducted at various stages of a project's lifecycle:

• Planning Stage: Identifies high-level safety impacts of route choices, access management, and general layout before significant design investments are made.
• Preliminary Design Stage: Focuses on alignment, cross-section, and intersection types. This is often the most cost-effective stage to implement changes.
• Detailed Design Stage: Reviews specific details like signage, pavement markings, lighting, clear zones, and landscaping.
• Construction Stage: Evaluates the safety of the temporary traffic control zones and checks the nearly completed project before it opens to the public.
• Existing Road RSA: Assesses operational roadways to identify safety deficiencies that have developed over time or were not apparent during design.

The RSA process involves assembling a multidisciplinary team (e.g., traffic engineers, law enforcement, human factors experts), conducting a pre-audit meeting, performing field reviews (often during both day and night), and producing a formal report of findings. The project owner must then provide a formal response documenting how each finding will be addressed.

By integrating EPDO analysis to quantify historical hazards, the HSM process to systematically manage improvements, human factors principles to design for actual user behavior, and RSAs to proactively catch design flaws, engineers can significantly reduce the societal burden of highway crashes.

Crash Rate Normalization

Raw crash counts do not provide a fair comparison of safety between different sites because they ignore exposure. A location with ten crashes per year and low volume is far more dangerous than a location with ten crashes per year and extremely high volume. Engineers normalize crash frequencies to account for traffic exposure.

Intersection Crash Rate (MEV)

Intersection safety is measured in crashes per Million Entering Vehicles (MEV). This metric aggregates all entering volumes from every approach leg: $R_{MEV} = \frac{A \times 1,000,000}{ADT_{\text{entering}} \times 365 \times N}$ where:

• $R_{MEV}$: Crash rate per million entering vehicles (crashes/MEV).
• $A$: Total number of crashes at the intersection during the study period.
• $ADT_{\text{entering}}$: Sum of average daily traffic entering the intersection across all approaches.
• $N$: Number of study years.

Roadway Segment Crash Rate (HMVM)

Roadway segments include a length component, and their exposure is measured in crashes per Hundred Million Vehicle Miles (HMVM) of travel: $R_{HMVM} = \frac{A \times 100,000,000}{ADT \times 365 \times L \times N}$ where:

• $R_{HMVM}$: Crash rate per hundred million vehicle miles (crashes/HMVM).
• $ADT$: Average daily traffic on the segment.
• $L$: Segment length (miles).
• $N$: Number of study years.

---

Collision Diagrams and Pattern Analysis

A collision diagram is a schematic drawing of an intersection or segment that uses standardized symbols to illustrate the path of each vehicle involved in a crash. It is an essential tool for identifying site-specific safety issues.

Common Collision Patterns and Engineering Countermeasures

Traffic engineers look for clusters of specific crash types to diagnose problems:

1. Rear-End Collisions:
   • Common Causes: Inadequate yellow change intervals or all-red clearance intervals, poor wet-weather pavement friction (skidding), sudden congestion, or obscured signal heads.
   • Countermeasures: Optimize signal timing, apply high-friction surface treatments, install advance warning signs ("Signal Ahead"), or add turn lanes to remove decelerating vehicles from the through-lanes.
2. Angle or Right-Type Collisions:
   • Common Causes: Restricted sight triangles, drivers running red lights/stop signs, or lack of turn control.
   • Countermeasures: Clear roadside vegetation/obstructions, install larger signal heads, adjust signal clearance intervals, or install a roundabout.
3. Left-Turn/Head-On Collisions:
   • Common Causes: Permissive left-turn phases with high opposing traffic volumes, poor visibility of oncoming traffic.
   • Countermeasures: Modify signal phasing to "protected-only" left turns, offset left-turn lanes to improve sight lines, or restrict left turns.

---

Hazard Identification: The Critical Rate Method

The Critical Rate Method is a rigorous statistical technique used to identify high-hazard locations. It compares the actual crash rate of a specific site ($R$) against a critical crash rate ($R_c$). If the actual rate exceeds the critical rate, there is a high probability (statistically significant) that the site is truly hazardous and not just experiencing random variations.

The critical crash rate is calculated as: $R_c = R_a + K \cdot \sqrt{\frac{R_a}{M}} + \frac{1}{2M}$ where:

• $R_c$: Critical crash rate (MEV for intersections, HMVM for segments).
• $R_a$: Average crash rate for similar facility types in the region.
• $K$: Statistical constant corresponding to the desired level of confidence.
• $M$: Traffic exposure (MEV for intersections, HMVM for segments).

Common values for the statistical constant $K$:

• $95%$ Confidence Level: $K = 1.645$
• $99%$ Confidence Level: $K = 2.326$
• $99.5%$ Confidence Level: $K = 2.576$

Exposure ($M$) is calculated as:

• For Intersections: $M = \frac{ADT_{\text{entering}} \times 365 \times N \times 10^{-6}}$
• For Segments: $M = \frac{ADT \times 365 \times L \times N \times 10^{-8}}$

---

Traffic Conflict Analysis

Crash data is reactive, requiring crashes to occur before action is taken. Traffic Conflict Analysis is a proactive safety method that observes and counts "near-misses" (conflicts) between road users.

Key Conflict Metrics

• Time-to-Collision (TTC): The time remaining before a collision would occur if the two vehicles continued on their present paths at their current speeds. Lower TTC values represent more severe conflicts.
• Post-Encroachment Time (PET): The time interval between when a first vehicle leaves a conflict area and a second vehicle enters it. Unlike TTC, PET does not require speed/trajectory extrapolation.

Conflict studies are valuable for evaluating safety improvements immediately after installation, without waiting years to gather statistically significant crash history.