A school crossing has a long queue of cars and an anxious group of pedestrians. Someone suggests keeping the vehicles green six seconds longer. The capacity calculation might look excellent—but where will the six seconds come from, and what happens to the people waiting to cross? Good traffic engineering asks both questions before changing the signals.
The gold standard of traffic engineering is a safe, evidence-based operating design whose movements, capacities, delays and real-world consequences can be checked by another competent analyst. This guide uses original calculations, changed-context tests, twenty detailed engineering controls and linked eduKate subject roots. Its numerical cases are teaching examples, never signal plans or permissions to modify live infrastructure.
Quick information: the distinction that prevents false confidence
| Measure | What it means | What it does not prove |
|---|---|---|
| Saturation flow | Discharge under defined continuous effective green | Actual capacity throughout a whole cycle |
| Effective green | Equivalent usable discharge time in a signal cycle | Free time that can be assigned without trade-offs |
| Capacity | Potential movement throughput under assumptions | A safe or queue-free junction |
| Demand-to-capacity ratio | Demand divided by the modelled capacity | A literal congestion percentage |
| Queue length | Waiting vehicles or distance under stated definition | The cause of congestion without further evidence |
| Travel time | Elapsed journey time over a defined boundary | Every passenger experiences the mean |
Two worked numerical cases
| Case | Assumptions | Calculation | Interpretation |
|---|---|---|---|
| Original movement | s = 1,800 veh/h/lane, g = 30 s, C = 90 s, v = 540 veh/h | c = 600 veh/h; v/c = 0.90 | Little nominal spare capacity under the simplified movement model |
| Reallocated effective green | s = 1,800, g = 36 s, C = 90 s, v = 540 | c = 720 veh/h; v/c = 0.75 | Movement ratio improves; other phases must be re-evaluated |
The equations apply to a simplified saturated movement and are not field estimates. The changed-context question matters more than the apparent arithmetic improvement: which conflicting movement or clearance obligation was affected? Check the entire signal plan, pedestrian access, actual queues and downstream network before concluding the change is desirable.
Twenty controls from observation to verified improvement
Define the junction problem
Chapters 1–5: What traffic engineering must actually achieve · Establish the study boundary and purpose · Collect turning movements and person movements · Understand the difference between flow, density and speed · Measure time and delay over the complete movement.
Calculate and test traffic capacity
Chapters 6–11: Describe the signal cycle and effective green · Calculate saturation flow before estimating capacity · Worked example: a signal movement has little spare capacity · Worked changed context: reallocating green time affects everyone · Identify the queue before diagnosing its cause · Use an oversaturation example without claiming exact field prediction.
Include every street user
Chapters 12–14: Treat pedestrian crossings as core network operations · Give buses and shared transport a fair accounting · Understand corridor progression and offsets.
Validate the engineering model
Chapters 15–18: Choose the simplest analysis tool that answers the question · Validate and calibrate a microsimulation separately · Safety cannot be traded away through one efficiency metric · Include weather, incidents and unusual operation.
Verify operational success
Chapters 19–20: Separate calculated capacity from operational success · Document implementation, monitoring and correction.
- What traffic engineering must actually achieve
- Establish the study boundary and purpose
- Collect turning movements and person movements
- Understand the difference between flow, density and speed
- Measure time and delay over the complete movement
- Describe the signal cycle and effective green
- Calculate saturation flow before estimating capacity
- Worked example: a signal movement has little spare capacity
- Worked changed context: reallocating green time affects everyone
- Identify the queue before diagnosing its cause
- Use an oversaturation example without claiming exact field prediction
- Treat pedestrian crossings as core network operations
- Give buses and shared transport a fair accounting
- Understand corridor progression and offsets
- Choose the simplest analysis tool that answers the question
- Validate and calibrate a microsimulation separately
- Safety cannot be traded away through one efficiency metric
- Include weather, incidents and unusual operation
- Separate calculated capacity from operational success
- Document implementation, monitoring and correction
CONTROL 1 OF 20 · Browse all controls
1. What traffic engineering must actually achieve
Traffic engineering is the practical design, analysis and management of movement on roads and intersections. The goal is not simply to push the maximum number of vehicles through a point. People walking, cycling, using public transport, driving or delivering goods all share parts of the system, while geometry, behaviour and signal timing create competing demands. A safe, understandable and dependable journey is the useful result.
A driver may encounter a short green period, yet extending it without examining pedestrian crossings and conflicting movements may relocate delay or create hazards. Traffic engineering begins with the whole operating context and all affected users. The broader network choices belong to Transport Planning; the appraisal of resources, congestion costs and economic trade-offs belongs to Transport Economics. A site-specific engineering decision must connect all three.
CONTROL 2 OF 20 · Browse all controls
2. Establish the study boundary and purpose
Define exactly which junctions, approaches, movements, people and time periods are being studied. A morning peak may be dominated by school arrivals while the evening peak includes more deliveries and different crossing patterns. A short corridor model that omits the upstream bottleneck can recommend a change that simply shifts the queue out of the model.
A useful study brief names the safety problem, baseline data, design constraints, network dependencies and intended outcome. For example, ‘reduce the probability of vehicles obstructing the school crossing during afternoon dismissal’ is a sharper question than ‘increase road efficiency’. The first can be tied to observable conditions; the second invites the analyst to choose a flattering measure after seeing results.
CONTROL 3 OF 20 · Browse all controls
3. Collect turning movements and person movements
Vehicle counts show how many vehicles turn or travel through an approach during a defined interval. They do not directly reveal how many people are carried, how many pedestrians cross or which routes are inaccessible. Record the count period, movement definition, vehicle classes, counting method and missing observations. Include bus occupancy or person flows where relevant and feasible.
Counts on an unusual wet day, school holiday or major-event day can misrepresent typical conditions. Compare multiple periods or explain why one period is appropriate. If a camera provides the data, account for sightlines, occlusion and privacy rules. A strong report distinguishes measured traffic counts from estimated demand and adjusted model inputs, and preserves a trace to the original observations.
CONTROL 4 OF 20 · Browse all controls
4. Understand the difference between flow, density and speed
Flow counts vehicles passing a location per unit time. Density describes vehicles occupying a length of road, and speed measures movement along it. In a simplified stationary traffic-stream relationship, flow equals density multiplied by space-mean speed. High flow is not a guarantee of high speed or safe operation, especially near saturation.
The relationship is conceptual when traffic varies quickly, lanes change and different vehicle types interact. Ask pupils to check the units: 20 vehicles per kilometre times 40 kilometres per hour gives 800 vehicles per hour under the simplified model. The numbers are fictional and do not describe an actual road. A count of 800 at one detector should not be mistaken for a permanent corridor capacity.
CONTROL 5 OF 20 · Browse all controls
5. Measure time and delay over the complete movement
Delay is additional time relative to a defined reference, and different methods measure different components. Stopped delay is not identical to control delay, total delay or door-to-door travel time. A claimed improvement must name which measure changed and how it was collected. A traffic-light algorithm can reduce stops without producing the same proportional reduction in total trip duration.
For an intersection, report the movements, time windows, sample sizes and variability. For a route, include upstream approaches and downstream merging where relevant. The Federal Highway Administration’s Traffic Analysis Tools programme provides guidance on choosing measures of effectiveness and suitable modelling methods. These are US technical resources, not automatically Singapore design approvals.
CONTROL 6 OF 20 · Browse all controls
6. Describe the signal cycle and effective green
A signal cycle is the interval over which the sequence of traffic movements repeats. The available green for a lane group is not simply a colour shown on a lamp: start-up and clearance effects matter. The effective green time is the equivalent time a movement can discharge at its saturation flow. The cycle and phase arrangement must account for conflicting vehicles, pedestrians and clearance intervals.
If a design claims that adding ten seconds of green creates free capacity, ask where those seconds came from. They may be taken from another movement and change queues, pedestrian service and safety. A phase diagram and an approved engineering basis are necessary for a real proposal. The FHWA Traffic Signal Timing Manual, Chapter 3 explains the engineering terminology and the effective-green model.
CONTROL 7 OF 20 · Browse all controls
7. Calculate saturation flow before estimating capacity
Saturation flow is an idealised discharge rate while queued traffic receives a continuous effective green under specified conditions. It is affected by lane use, vehicle mix, grades, turns, conflicting flows and local operating behaviour. It must not be copied from an unrelated location merely because a spreadsheet contains a default.
For a queue whose stable departure headway is 2.5 seconds per vehicle, the idealised saturation flow is 3,600 ÷ 2.5 = 1,440 vehicles per hour per lane. This is a mathematical example, not a measured Singapore standard. Actual site analysis needs observed headways, geometry, classification and an appropriate technical method. A small headway difference can materially change a capacity result.
CONTROL 8 OF 20 · Browse all controls
8. Worked example: a signal movement has little spare capacity
Consider a fictional intersection approach with an assumed saturation flow of 1,800 vehicles per hour per lane, a 90-second cycle and 30 seconds of effective green. The simplified capacity expression is c = s × g / C. It gives 1,800 × 30 / 90 = 600 vehicles per hour per lane. If observed demand is 540 vehicles per hour, the volume-to-capacity ratio is 540/600 = 0.90.
This 0.90 result says the demand is 90% of the assumed movement capacity under that approximation. It does not directly tell us the queue’s 95th percentile, exact waiting time, crash risk or pedestrian delay. All numbers are invented to demonstrate the equation. Check the mathematical assumptions, phase interactions and official procedures before using a similar method for actual signal design.
CONTROL 9 OF 20 · Browse all controls
9. Worked changed context: reallocating green time affects everyone
Suppose an analyst raises the hypothetical movement’s effective green from 30 to 36 seconds while keeping a 90-second cycle and saturation flow of 1,800 vehicles per hour per lane. The nominal movement capacity rises to 1,800 × 36/90 = 720 vehicles per hour per lane. At demand 540, its volume-to-capacity ratio falls to 0.75. Those results are mathematically coherent.
But the extra six effective-green seconds cannot appear from nowhere. In a real signal plan, changing phase splits, clearances or cycle length may reduce service for another approach or crossing. The improved movement’s isolated ratio cannot prove that the intersection as a whole works better. Reassess pedestrian safety, queues, network progression and other movements. This changed-context check prevents a correct formula from supporting an unjustified decision.
CONTROL 10 OF 20 · Browse all controls
10. Identify the queue before diagnosing its cause
A queue can result from normal red time, a downstream blockage, turning conflicts, insufficient effective green, irregular arrivals or an incident. Vehicles waiting at a stop line are an observation, not a diagnosis. Measure when the queue forms, how far it extends, what blocks discharge and whether it clears in each cycle.
Upstream spillback can obstruct a different intersection. A camera aimed only at the stop line may miss this. Report queue length in explicit units and note whether it is a mean, maximum observation or a percentile. Queue metres cannot be reliably inferred from vehicle count without assumptions about vehicle length, spacing and composition. A responsible engineering account names those assumptions rather than presenting a single precise-looking queue length.
CONTROL 11 OF 20 · Browse all controls
11. Use an oversaturation example without claiming exact field prediction
Assume, for illustration, a continuous arrival stream of 600 vehicles an hour and a simplified service rate of 480 vehicles an hour for the same approach over a sustained interval. Arrivals exceed departures by 120 vehicles an hour, or two vehicles each minute. After 15 minutes, the net accumulated backlog is about 30 vehicles if rates remain constant and nothing else changes.
This is a fluid-flow approximation, not a physical simulation of signal cycles. Real arrival bursts, lane blocking, capacity drops, signal changes and turns can alter queues. The calculation is useful because it identifies a structurally unstable sustained demand-service relationship. A red-time queue that clears each cycle should not be confused with one that accumulates without recovery.
CONTROL 12 OF 20 · Browse all controls
12. Treat pedestrian crossings as core network operations
Pedestrian signals, crossings, refuge points and approach visibility influence mobility and safety. A plan that shortens every vehicle delay yet makes a crossing unusable for a slower pedestrian is not a complete improvement. Engineering review must consider all users and the required legal and technical rules for the jurisdiction.
Observe whether people can reach the crossing, wait safely, see the indication and finish through an appropriate protected movement. Pedestrian demand may peak at school times even if vehicle flow peaks elsewhere. Do not derive real pedestrian clearance settings from a generic internet calculation; a qualified designer must apply the relevant local standards, site constraints and safeguarding requirements.
CONTROL 13 OF 20 · Browse all controls
13. Give buses and shared transport a fair accounting
A traffic stream of 30 cars and one bus is not simply 31 equivalent people-movement units. Vehicle count is an engineering measure, but person throughput and equitable access may be better policy measures for some questions. A bus carrying many passengers can offer substantial corridor capacity even when it occupies more road space than one car.
Bus priority may involve signal control, lanes, stops or junction geometry. Evaluate travel-time reliability for bus passengers alongside effects on conflicting movements. The traffic model should identify bus dwell times and stop locations. Labelled vehicle-to-passenger assumptions prevent an impressive vehicle-delay chart from concealing a worsening journey for more people.
CONTROL 14 OF 20 · Browse all controls
14. Understand corridor progression and offsets
Signal coordination can reduce unnecessary stopping when vehicles travel along a corridor at prevailing speeds. Offset is the temporal relationship between signal cycles at successive junctions. It is often designed around a specified direction, time period and traffic pattern, so an offset cannot create simultaneous green waves for every movement under all conditions.
A school crossing, bus-priority intervention or new junction demand can require reconsideration of the coordination plan. Review both directions and side streets, not merely the time selected for a promotional video. Modelling should consider how vehicles disperse after each junction, whether a downstream queue blocks entry and whether pedestrians have safe, predictable access.
CONTROL 15 OF 20 · Browse all controls
15. Choose the simplest analysis tool that answers the question
A spreadsheet can sometimes check flow and capacity for a well-defined movement. More complex networks may require analytical models, mesoscopic tools, microscopic simulation or direct field experiments. Greater software detail does not automatically create more accurate evidence. Complex models bring calibration, assumptions, data and random-seed uncertainty.
Select a tool after specifying the question, geographic coverage, required measures and availability of reliable inputs. FHWA’s Traffic Analysis Toolbox includes a decision-support methodology and 2019 updated microsimulation guidance. Match the model to the decision rather than choosing an elaborate animation because it looks persuasive.
CONTROL 16 OF 20 · Browse all controls
16. Validate and calibrate a microsimulation separately
Calibration adjusts model parameters so simulated behaviour reasonably matches selected observed conditions. Validation uses appropriate independent evidence to evaluate whether the model can represent the relevant setting or support the intended inference. A simulation configured until one travel-time value matches a survey has not necessarily reproduced queues, demand, lane changes or the conditions of a different day.
Use multiple relevant measures and intervals where appropriate, document parameter changes and retain independent checks. If the model uses random draws, report repetitions and variation rather than selecting the best run. The FHWA 2019 microsimulation guidance gives a structured approach. The authority of the process comes from disciplined evidence, not from the brand of software used.
CONTROL 17 OF 20 · Browse all controls
17. Safety cannot be traded away through one efficiency metric
A junction with fewer seconds of mean delay may still expose turning vehicles and pedestrians to greater conflict. Before implementing changes, review sightlines, speeds, movements, vulnerable users and applicable standards. Safety considerations need their own methods and qualified judgement, not simply a cost–benefit claim.
When risk is uncertain, report the uncertainty and consider protective alternatives. A change that improves a car queue while diverting walking routes into unsafe crossings is not a clean gain. Use Risk Management to structure uncertain hazards, Systems Engineering to trace requirements and tests, and Urban Planning to connect street space with its intended community use.
CONTROL 18 OF 20 · Browse all controls
18. Include weather, incidents and unusual operation
An intersection operates under changing conditions. Heavy rain, faulty detectors, road works, a stalled vehicle and a special event may change demand and discharge behaviour. A plan validated only on a quiet dry weekday may misrepresent the periods when errors matter most. Emergency procedures must be governed by the proper authorities.
Choose a realistic range of operating scenarios and test what the controller and users experience when expected signals or data are unavailable. For a monitored system, keep time-stamped incident data, repair records and a defined escalation path. Avoid reporting planned timings as observations of actual intersection performance.
CONTROL 19 OF 20 · Browse all controls
19. Separate calculated capacity from operational success
An increase in theoretical approach capacity does not prove shorter travel times, improved reliability, lower crashes or better accessibility. The additional capacity may be unreachable due to a downstream bottleneck, or it may attract demand that changes the network. Success must be defined with the affected people and full route in mind.
Use a balanced set of measures such as queue spillback frequency, travel-time distribution, pedestrian service, bus delay and observed conflicts, according to the decision. The analyst should state which measures are forecasts, which are observed and which remain unmeasured. Connect those findings to Transport Economics before making a large public investment claim.
CONTROL 20 OF 20 · Browse all controls
20. Document implementation, monitoring and correction
A signal change is an intervention with a responsible owner, approved controller settings, baseline measurements, deployment time, testing protocol and rollback conditions. A successful launch should not be inferred from one green dashboard indicator. Monitor performance after implementation and compare relevant time periods, acknowledging changes outside the project.
Retain the version of the timing plan that generated each result. Investigate unexpected spillback, accessibility complaints and worsened side-street conditions. Publish corrective decisions and their verification. For a student, this is the same habit as checking the worked method in a Mathematics problem and transferring it to a new case: every claimed improvement must survive a second explanation, not simply an attractive number.
Practice check: diagnose the tempting wrong answer
A student uses the right equation, c = s × g/C, obtains a plausible capacity and declares that the intersection is safe. What is missing? Answer: a capacity calculation for one movement does not evaluate conflicting movements, pedestrian crossing, phase clearances, queue propagation or safety. The method answered a narrower question. Before advancing, write the omitted assumptions and a test that would detect a harmful transfer of delay.
Operational traffic-engineering audit
| Review item | Required evidence | Common failure |
|---|---|---|
| Study scope | Junctions, movements, users, hours and comparison period | A corridor truncated before the bottleneck |
| Counts and field data | Method, dates, turning movements, unusual conditions | Simulation assumptions presented as observed flows |
| Signal model | Cycle, phases, clearance, effective green and demand | Extra green created without phase trade-off |
| Safety | Vulnerable users, crossings, hazards, approved technical review | Only vehicle delay considered |
| Simulation | Inputs, calibration, independent checks and variability | One favourable animation taken as proof |
| After-change review | Installed settings, before/after observations, rollback plan | Improvement declared before field evaluation |
Frequently asked questions
Is traffic engineering the same as transport planning?
No. Traffic engineering focuses on detailed operations and design of traffic movements, especially roads and intersections; transport planning addresses wider access, land use and network choices. They interact.
Does a 0.90 volume-to-capacity ratio mean exactly 90% congestion?
No. It is a ratio of assumed demand to calculated capacity for the defined movement and period. It does not directly specify travel time, queue length or safety.
Can extending green time fix congestion?
Sometimes for one movement, but it can take capacity or safe crossing time from another. The whole signal plan and network must be checked.
Does microsimulation predict the future precisely?
No. Results depend on demand, behaviour, calibration, network boundaries and uncertainty. Different seeds or scenarios can yield different outputs.
Why count people as well as vehicles?
Vehicle throughput does not represent the number of people reaching destinations; buses and cars may carry different passenger loads.
What should an independent reviewer receive?
Baseline and test data, assumptions, measures, calibrated model or calculations, alternatives, safety review, accepted plan version and post-change observations.
Authoritative technical resources and boundaries
The U.S. Federal Highway Administration Traffic Signal Timing Manual, Chapter 3 defines saturation flow, effective green, capacity and delay concepts. The FHWA Traffic Analysis Toolbox offers traffic analysis selection guidance and the 2019 microsimulation update discusses modelling process, calibration and analysis. These are authoritative methodology sources in their contexts; design and signal operation in Singapore require applicable local rules and qualified professional approval.
All queue, saturation, effective-green and vehicle-count examples are invented for teaching. They are not measurements of a named road, site certification or traffic controller configuration. This article distinguishes analytical capacity from operational safety and tested whole-journey performance.
The connected eduKate Gold Standard subject network
Transport Planning · Transport Economics · Transportation · Engineering · Systems Engineering · Urban Planning · Infrastructure · Risk Management · Statistics · Evidence-Based Thinking · Geography · Education
For learners, start with units, ratios and the meaning of a result, then test a changed context. The immutable Secondary 1 Mathematics Tutor Clementi reference demonstrates the underlying teaching habit: explain each move, diagnose the mistaken assumption and confirm that understanding transfers. It does not supply road-engineering design values.
Return to all controls · Continue with Transport Planning · Continue with Transport Economics