Traffic Engineering: How Traffic Flow, Signals, and Intersections Are Analyzed

Traffic engineers turn observations and data into decisions. A typical assignment may begin with a complaint such as “this intersection is unsafe” or “traffic backs up every afternoon.” Those statements describe a concern, but they do not yet define the engineering problem.

The engineer first visits the location or reviews current video. Are drivers unable to see oncoming traffic? Does the queue block a turn lane? Are pedestrians waiting through multiple cycles? Is the congestion limited to school dismissal? Does a downstream signal prevent traffic from leaving the intersection? Field context often changes the entire direction of the study.

Common traffic engineering assignments

• Traffic volume studies: measuring daily traffic, peak-hour demand, turning movements, trucks, pedestrians, and bicycles.
• Intersection evaluations: analyzing control delay, queues, lane use, sight distance, turn-lane storage, and control type.
• Signal timing: adjusting cycles, phases, splits, offsets, clearance intervals, detection, and pedestrian timing.
• Traffic impact studies: estimating how a proposed development changes traffic demand, access, queues, and nearby operations.
• Safety studies: diagnosing crash patterns and selecting countermeasures that address the actual contributing factors.
• Traffic control plans: designing temporary signs, channelization, detours, and lane closures for construction or special events.
• Access management: evaluating driveways, medians, turn restrictions, spacing, and internal circulation along a corridor.

The deliverable may be a short memorandum, a full traffic impact analysis, a signal timing plan, construction drawings, or recommendations for a capital project. Regardless of format, the strongest work connects every recommendation to observed conditions and documented evidence.

Transportation engineering is the broader discipline. It includes transportation planning, geometric roadway design, pavement design, transit, freight, airports, rail, ports, operations, safety, technology, economics, and policy.

Traffic engineering sits inside that broader field and concentrates on operations and control. It asks how the system performs now, how it will perform under a proposed change, and which operational or geometric improvement best addresses the identified problem.

Traffic engineering vs. highway design

Highway design focuses on the physical roadway: horizontal and vertical alignment, lane and shoulder widths, medians, cross slopes, roadside features, and sight distance. Traffic engineering focuses on how users move through that geometry. The two disciplines overlap when traffic demand determines the number of lanes, turn-lane length, intersection layout, or signal placement.

Traffic engineering vs. transportation planning

Transportation planning often examines future networks, land use, travel demand, funding priorities, and long-range policy. Traffic engineering typically works at the site, intersection, corridor, or network-operations level. A planning model may identify a future bottleneck; a traffic engineering study determines how that bottleneck functions and what practical improvement should be implemented.

1. Start with the problem, not the preferred solution

Requests often arrive as solutions: “install a signal,” “add a lane,” or “put in a stop sign.” The engineer must step back and identify the underlying issue. A reported delay problem might actually be short turn-lane storage. A reported speeding problem might be created by roadway geometry and low roadside activity. A crash pattern may be caused by limited sight distance rather than insufficient control.

2. Use representative data

Traffic conditions change by hour, day, season, school schedule, weather, construction activity, and nearby events. Data collected during an abnormal period can produce a precise but misleading answer. The engineer should document when data was collected and whether it represents the condition being studied.

3. Analyze the system, not one isolated point

Intersections and road segments influence one another. A signal may discharge more traffic than the next intersection can accept. A driveway queue may block a public street. A ramp queue may extend onto a freeway. Traffic engineering becomes most valuable when it identifies these connections rather than treating each location as independent.

4. Balance mobility and safety

Faster movement is not always safer or more useful. Wider crossings, longer cycles, permissive turns, and additional lanes may reduce vehicle delay while increasing pedestrian exposure or crash severity. A complete analysis considers the type of users, operating speed, conflict points, visibility, and the consequences of a mistake.

5. Design for people, not only passenger cars

Trucks need turning space. Buses need stopping and merging space. Pedestrians need visible, direct crossings and enough time. Cyclists need predictable paths through conflict areas. Emergency vehicles may need priority. A recommendation that works only for the average passenger car is incomplete.

6. Check the recommendation in the field

Analysis predicts performance; field observation confirms it. After implementation, engineers should review whether queues actually shortened, speeds changed, users understood the new control, and unintended problems appeared. Before-and-after monitoring is how the profession learns whether a solution worked.

Traffic conditions cannot be understood from volume alone. Two intersections can serve the same number of vehicles and perform very differently because of lane arrangement, signal timing, turning percentages, pedestrians, trucks, driveway activity, or downstream congestion.

Traffic volume and flow rate

Volume is the number of users counted during a stated period. Flow rate expresses that demand as an equivalent hourly rate. Turning movement counts are especially important because left, through, and right movements use different lanes and conflict with different users.

Speed and travel time

Spot-speed studies describe speeds at a location. Travel-time studies describe how long it takes to move through a corridor and how much that time varies. Reliability matters because a route that takes 15 minutes on one day and 35 minutes on another is harder to plan around than a consistently slower route.

Delay

Delay is the additional time users experience compared with a reference condition. At intersections, control delay is a common measure. It captures the effect of slowing, stopping, waiting, and accelerating because of traffic control.

Queue length

Queue length is often more important to design than average delay. A turn lane can have acceptable average delay but still overflow into a through lane during several cycles. Engineers compare expected queues with available storage and check what the queue blocks when it grows.

Volume-to-capacity ratio

The volume-to-capacity ratio compares demand with the estimated ability of a movement or facility to serve that demand. A value near 1.0 indicates that demand is approaching estimated capacity. Values above 1.0 suggest that the queue is likely to grow unless demand drops or additional discharge capacity becomes available.

Level of service

Level of service, or LOS, summarizes operating quality using letters A through F. The performance measure behind the grade depends on the facility. Signalized intersections commonly use control delay; freeway segments commonly use density. LOS is useful for communication, but it should never replace the underlying delay, queue, speed, density, safety, and multimodal results.

A traffic study should collect the minimum data needed to answer the actual question—without overlooking conditions that could change the conclusion. The following study types are common because each reveals a different part of the operating picture.

Turning movement counts

Turning movement counts record left, through, and right movements by approach, usually in 15-minute intervals. Good counts also identify pedestrians, bicycles, and heavy vehicles when those users affect the analysis. These counts form the basis for intersection capacity, signal timing, and turn-lane evaluation.

Daily and classification counts

Pneumatic tubes, radar, cameras, and permanent stations can measure traffic over one or more days. The data shows daily variation, peak periods, directional patterns, and sometimes vehicle class. A short count may need adjustment when annual average daily traffic or seasonal conditions matter.

Speed studies

Speed studies help evaluate operating conditions, speed-limit questions, stopping requirements, and traffic-calming needs. Engineers should distinguish free-flow speeds from speeds suppressed by congestion, signals, or a temporary queue.

Queue and travel-time studies

Queue observations show whether vehicles clear each cycle and whether storage is sufficient. Travel-time runs, probe data, or connected data can reveal where delay occurs along a corridor and whether signal coordination is working.

Crash data and conflict observations

Crash records provide location, severity, type, time, and contributing information, but they may not explain the physical mechanism by themselves. Collision diagrams, site photographs, video, sight-distance checks, and observations of near conflicts help the engineer connect the record to roadway conditions.

For a deeper look at counts, sensors, AADT, peak-hour factors, and field equipment, see Traffic Monitoring.

Traffic flow theory connects three basic variables: flow, density, and speed. The relationship is commonly written as:

When density is low, vehicles can usually travel near free-flow speed. As more vehicles enter the road, flow increases, but interactions between drivers also increase. Near capacity, small disturbances—braking, merging, lane changes, a truck on a grade, or a short signal interruption—can cause speeds to drop sharply.

Once traffic enters congested operation, adding more vehicles does not necessarily increase the number served. Speed can collapse, gaps become inconsistent, and queues propagate upstream. This explains why congestion sometimes appears without a crash or lane closure: demand has reached a condition where ordinary disturbances no longer dissipate.

Learn more in the dedicated Traffic Flow Theory guide.

Capacity is the maximum sustainable rate at which a facility or movement can serve traffic under stated conditions. It is not a universal number. Lane width, grade, heavy vehicles, parking activity, turning movements, signal timing, pedestrian conflicts, merges, and driver behavior can all affect the estimate.

Why queues matter more than averages

Average delay can hide the most damaging condition. A left-turn movement may operate acceptably for most cycles but occasionally produce a queue long enough to block the through lane. Once that happens, the intersection no longer behaves like the isolated movement represented in the calculation.

Engineers therefore compare modeled or observed queue lengths with available storage. They also identify what lies upstream: another signal, a railroad crossing, a freeway ramp, a school entrance, or a major driveway. Queue spillback is a network problem, not only an intersection statistic.

How LOS should be used

LOS is best used as one summary of operating quality. It can help agencies compare alternatives and communicate results, but it should be presented with the underlying performance. A recommendation is much stronger when it explains that a movement has high delay, a volume-to-capacity ratio near 1.0, a 95th-percentile queue longer than storage, and a documented crash or access issue.

A traffic signal assigns right-of-way to movements that cannot proceed safely at the same time. The challenge is to serve competing demands without creating excessive delay, unsafe clearances, confusing transitions, or queues that interfere with nearby locations.

Cycle length

The cycle length is the time required for the signal to complete one sequence of phases. Short cycles serve movements more frequently but lose a larger share of time to phase changes. Long cycles can provide more capacity to major movements but increase waiting time and may create long pedestrian delays.

Phases and splits

A phase gives right-of-way to one or more compatible movements. The split is the portion of the cycle assigned to that phase. Protected and permissive turns, pedestrian intervals, transit priority, and emergency preemption can make the phase sequence more complex.

Offsets and progression

The offset establishes the timing relationship between signals along a corridor. Good coordination can create a progression band that allows platoons to move through several intersections with fewer stops. Poorly chosen offsets can shift delay from one location to another or cause a downstream queue to block an upstream signal.

Clearance and pedestrian intervals

Yellow and all-red intervals support a safe transition between conflicting movements. Pedestrian timing must provide a usable walk indication and enough clearance time for the crossing. These are safety requirements, not spare time to be removed whenever vehicle demand increases.

This simplified relationship shows why signal timing changes capacity: a lane group can discharge traffic only during its effective green. Real analysis also accounts for lost time, turning movements, heavy vehicles, lane use, pedestrians, and local conditions.

Continue with Signal Timing and Phasing and Traffic Signals.

Intersections concentrate conflict into a small area. Drivers cross opposing traffic, turn across pedestrians, merge into gaps, respond to signs or signals, and make decisions while watching several sources of information. That is why intersection analysis combines operations, geometry, visibility, human factors, and safety.

Stop-controlled intersections

At a two-way stop, the major street moves freely while minor-street drivers wait for acceptable gaps. The engineer reviews minor-street delay, queue length, sight distance, speed, turning demand, and the availability of gaps. An all-way stop may be appropriate in some balanced or safety-related conditions, but it should not be installed merely because drivers request equal stopping.

Signalized intersections

Signal analysis examines lane groups, phasing, detection, pedestrian service, queue storage, and coordination with nearby signals. Meeting a traffic signal warrant is part of the evaluation; it does not automatically mean a signal is the best treatment.

Roundabouts

Roundabouts reduce crossing and opposing-turn conflicts by moving traffic in one direction around a central island. Their performance depends on entry geometry, circulating flow, approach speed, pedestrian treatments, and the design vehicle. A poorly designed roundabout can create path overlap or excessive entry speed even when a software model reports acceptable delay.

Turn lanes and access

Turn lanes separate slowing or waiting vehicles from through traffic, but they require adequate deceleration and storage. Driveway spacing, median openings, left-turn access, and internal circulation can determine whether a corridor remains functional as development increases.

See Intersection Design for the geometric side of these decisions.

A traffic impact study evaluates how a proposed development or land-use change may affect surrounding transportation facilities. The study may address a subdivision, apartment complex, warehouse, school, retail center, industrial project, mixed-use development, or major site expansion.

1. Agree on the study scope

The applicant and reviewing agency identify the intersections, road segments, analysis years, peak periods, planned projects, and assumptions to include. A clear scope prevents major disagreements after the analysis is complete.

2. Document existing conditions

The engineer gathers traffic counts, signal timing, lane geometry, speed information, crash history, pedestrian activity, access locations, planned roadway projects, and field observations.

3. Estimate future traffic

The analysis considers background growth, approved developments, and trips generated by the proposed project. The engineer documents trip generation, internal capture, pass-by trips, directional distribution, and route assignment rather than presenting these assumptions as unexplained software inputs.

4. Compare scenarios

Existing, no-build, and build conditions are analyzed so the development’s effect can be separated from background conditions. Depending on the project, the study may examine delay, queues, turn-lane needs, driveway operation, signal warrants, sight distance, pedestrians, trucks, and internal circulation.

5. Develop practical mitigation

Mitigation may include access changes, turn lanes, signal modifications, crosswalk improvements, driveway consolidation, median treatments, or participation in a larger corridor project. A recommendation should be proportionate to the identified impact and feasible within the project’s right-of-way, schedule, and approval process.

A crash total tells the engineer that events occurred; it does not explain why. Useful safety analysis looks for repeatable patterns involving crash type, severity, direction, time, weather, lighting, speed, geometry, visibility, and user behavior.

Rear-end patterns

Rear-end crashes may be associated with unexpected queues, limited signal visibility, large speed differences, short turn-lane storage, or inconsistent stopping. The response could involve advance warning, signal timing, visibility, speed management, or queue reduction depending on the mechanism.

Angle and left-turn patterns

Angle crashes may indicate gap-selection difficulty, sight restrictions, signal compliance issues, or confusing control. Left-turn crashes may support protected phasing, access changes, turn restrictions, or geometric improvements, but each treatment creates tradeoffs that must be evaluated.

Pedestrian and bicycle risk

Engineers review crossing distance, speed, lighting, visibility, refuge, signal timing, turning conflicts, and whether the expected path is direct and understandable. A marked crosswalk by itself does not correct every crossing problem.

Speed and crash severity

Higher operating speed increases stopping distance and the energy involved in a collision. Speed management can include geometric design, lane configuration, roadside context, signal progression, traffic calming, enforcement support, and clear transitions between roadway environments.

Continue with Road Safety, Accident Analysis, and Traffic Calming Measures.

Consider a signalized intersection where drivers report that the eastbound left-turn queue blocks the adjacent through lane during the morning peak. The purpose of this simplified example is to show how several measures work together. It is not a substitute for a full HCM analysis.

Step 1: Check how concentrated the demand is

The four 15-minute left-turn counts are 125, 140, 165, and 145 vehicles. The hourly volume is therefore 575 vehicles. The peak-hour factor is:

A PHF of 0.87 shows that demand is noticeably concentrated in the busiest 15 minutes. Using only the hourly average would understate the short period when the queue is most likely to overflow.

Step 2: Estimate the available service rate

Assume one exclusive left-turn lane, a simplified saturation flow rate of 1,900 vehicles per hour of green, 30 seconds of effective green, and a 90-second cycle. The simplified capacity is:

The demand-to-capacity ratio is approximately \(575/633 = 0.91\). That result suggests the movement is operating close enough to its estimated capacity that arrival variation, heavy vehicles, lost time, or downstream interference can create substantial queues.

Step 3: Compare the queue with physical storage

Field video shows a recurring maximum queue of about 10 passenger-car equivalents. At roughly 25 feet per queued vehicle including spacing, the queue occupies about 250 feet. The painted storage length is only 180 feet, so the queue extends about 70 feet into the through lane during the busiest cycles.

Step 4: Identify the real problem

The key issue is not simply “LOS” or “high volume.” The left-turn lane lacks storage for the concentrated peak demand, and the overflow interferes with through traffic. That interference can reduce through capacity and create sudden lane changes upstream.

Step 5: Evaluate alternatives

The engineer could test additional left-turn green time, a longer storage bay, revised phasing, or access changes that reduce turning demand. Increasing green time may affect other approaches and pedestrian service. Extending the bay may require right-of-way or remove another roadway feature. The recommendation should compare these tradeoffs and confirm the result using the governing analysis method.

Traffic engineers use spreadsheets, capacity-analysis programs, signal optimization tools, GIS, crash-analysis systems, and microscopic or mesoscopic simulation. The correct tool depends on the question and the complexity of interaction between facilities.

Spreadsheets and transparent calculations

Spreadsheets are useful for count summaries, growth calculations, warrant checks, queue-storage checks, and clear supporting calculations. They are often the best choice when the method is simple and reviewers need to follow every assumption.

Capacity and signal analysis

HCM-based software estimates delay, capacity, queues, density, and level of service for applicable facility types. Signal tools help evaluate cycles, splits, offsets, coordination, and progression. These programs require careful lane configuration, phasing, timing, and demand inputs.

Traffic simulation

Simulation is useful when individual vehicle interactions, closely spaced intersections, complex ramps, oversaturated queues, transit operations, or unusual lane behavior cannot be represented adequately by a simpler method. More detail does not automatically mean more accuracy. The model must be calibrated to observed travel times, queues, speeds, and discharge behavior.

GIS and spatial safety analysis

GIS helps map crashes, roadway features, access points, transit stops, and corridor patterns. Spatial analysis can reveal clusters and network context that are difficult to see in a spreadsheet.

See Traffic Engineering Software for a focused overview.

Although project requirements differ, the following workflow keeps analysis tied to the real decision.

1. Define the concern. Describe the reported problem, affected users, location, time period, and consequence.
2. Observe the site. Review geometry, control, visibility, queues, access, pedestrian behavior, truck movements, and nearby constraints.
3. Set the study scope. Identify the facilities, time periods, scenarios, data, and governing methods needed to answer the question.
4. Collect representative data. Document counts, speeds, queues, travel times, crashes, timing, geometry, and unusual conditions.
5. Diagnose existing conditions. Connect calculated results with what occurs in the field.
6. Forecast the relevant future condition. Include growth, development, planned projects, and changes in access or control when required.
7. Develop more than one alternative. Include operational, geometric, safety, access, and demand-management options where appropriate.
8. Evaluate tradeoffs. Consider safety, delay, queues, pedestrians, cyclists, transit, freight, cost, right-of-way, construction, and maintenance.
9. Document the recommendation. State assumptions, methods, limitations, and why the preferred alternative solves the identified problem.
10. Monitor after implementation. Confirm that users understand the change and that the expected operational or safety benefit occurs.

Using a count that does not represent normal demand

A count collected during a school holiday, lane closure, severe storm, special event, or nearby construction may not represent the condition being designed for. Record the context and collect supplemental data when necessary.

Optimizing only vehicle delay

A change that improves one vehicle metric can increase crossing distance, speed, conflict exposure, or delay for another user. Report multimodal and safety consequences alongside vehicle operations.

Ignoring queue spillback

An isolated intersection result may look acceptable even though its queue blocks a driveway, ramp, railroad crossing, or upstream signal. Compare queue estimates with actual storage and inspect the entire influence area.

Treating a warrant as a design decision

A warrant helps identify when a treatment may be considered. It does not remove the need for an engineering study, alternatives analysis, and judgment about whether the treatment is appropriate.

Trusting default software inputs

Defaults may not represent local saturation flow, driver behavior, lane use, walking speed, heavy vehicles, or signal operation. Important assumptions should be verified and documented.

Designing for the average vehicle only

The governing design vehicle may be a bus, fire apparatus, delivery truck, or combination vehicle. Off-tracking, curb-return geometry, lane encroachment, and storage must be checked where those vehicles operate.

Traffic engineers commonly begin with a civil engineering degree and coursework in transportation, statistics, probability, highway design, and engineering economics. Entry-level work often includes traffic counts, field observations, data processing, capacity analysis, signal timing support, traffic control plans, and report preparation.

Strong traffic engineers develop both technical and observational skills. They learn to question unusual model results, recognize when a count is unrepresentative, explain a queue in physical terms, and communicate tradeoffs to reviewers and the public.

Skills that matter early

• Reading roadway and traffic-control plans
• Working with traffic count and crash data
• Understanding flow, capacity, delay, and queueing
• Writing clear assumptions and recommendations
• Performing field observations safely and systematically
• Learning the governing agency’s standards and review process
• Explaining technical results without hiding behind software output

Professional licensure requirements vary by jurisdiction and responsibility. Engineers preparing or approving final designs should follow applicable licensing laws, agency criteria, and organizational quality-control procedures.

AADT: Annual average daily traffic, representing average daily traffic over a year.

Approach: The part of an intersection used by traffic entering from one direction.

Control delay: Delay caused by deceleration, queueing, stopping, and acceleration associated with traffic control.

Cycle length: The time required for a traffic signal to complete one full sequence of phases.

Density: The number of vehicles occupying a unit length of roadway, often expressed per lane.

Effective green: The portion of a cycle during which a movement effectively discharges traffic.

Level of service: A letter-grade description of operating quality based on a facility-specific performance measure.

Peak-hour factor: A measure of how evenly traffic demand is distributed within the peak hour.

Phase: A signal interval that assigns right-of-way to one or more compatible movements.

Queue spillback: A queue extending far enough upstream to block another lane, driveway, ramp, or intersection.

Saturation flow: The rate at which a stable queue can discharge while receiving effective green under stated conditions.

Volume-to-capacity ratio: The ratio of demand to estimated capacity for a movement or facility.