Definitions and Discussion: 16-Civ-B10 Traffic Engineering, Undated paper, Q1

Question 1: Definitions and Discussion (5 × 4 = 20 marks)

Question text not reproduced: the examination questions are © Engineers and Geoscientists BC. Open the official past paper (linked at the top of this page) to read the question, then follow the worked solution below.

Part (a) — Leading versus lagging protected phase. A protected left-turn phase is an interval during which left-turning vehicles receive a green arrow and the opposing through movement is held at red, so the turn is made without any gap-acceptance decision. The question is only when in the signal sequence that protected interval is placed relative to the adjacent through movement. In a leading protected phase the green arrow is displayed before the parallel through green: the sequence is left-arrow, then the through green for both directions. In a lagging protected phase the arrow follows the through green, so both directions run through movements first and the left turn is cleared at the end of the phase group.

The engineering trade-offs are real and pull in opposite directions. Leading lefts are the North American default because they exploit the queue that has accumulated during red — the left-turn bay is full at the instant the arrow appears, so the saturation flow is achieved almost immediately and the arrow can be short. They also match driver expectancy: a driver stopped at a red arrow expects to be released first. Their weakness is the left-turn trap when one direction runs protected-permissive and the opposing direction lags; a driver turning permissively on a circular green may not realise that the opposing through movement is still running. Lagging lefts, by contrast, are the natural choice when a leading arrow would strand pedestrians, when the through movement is being coordinated in a progression band (the through band is best placed at the start of the phase so the platoon is not held), or when a lead-lag arrangement is used deliberately to shift the green band between two adjacent intersections. Lagging lefts also permit a longer effective through green for a platoon arriving early. In Canadian practice, MUTCDC display rules and the associated signal-head configuration must be checked before a lag is introduced, because the left-turn trap is a documented crash mechanism.

Part (b) — Time mean speed versus space mean speed. Both are averages of individual vehicle speeds; they differ in the sampling frame, and that difference is not a technicality. Time mean speed $u_t$ is the arithmetic mean of the spot speeds of vehicles passing a fixed point during a stated period, $u_t = \frac{1}{n}\sum u_i$. It is what a radar gun or a loop detector measures. Space mean speed $u_s$ is the mean speed of the vehicles occupying a stated length of road at an instant, and is computed as the harmonic mean of the individual speeds, or equivalently as the section length divided by the mean travel time:

$$u_s=\frac{n}{\sum \dfrac{1}{u_i}}=\frac{d}{\dfrac{1}{n}\sum t_i}$$

Because the harmonic mean can never exceed the arithmetic mean, $u_s \le u_t$ always, with equality only when every vehicle travels at the same speed. The two are related exactly by

$$u_t=u_s+\frac{\sigma_s^{2}}{u_s}$$

where $\sigma_s^{2}$ is the variance of the space-mean-speed distribution. The practical significance is that only space mean speed belongs in the fundamental flow relation $q = k\,u_s$. A point measurement over-samples fast vehicles — more of them pass the detector in a given period — so substituting $u_t$ into $q = ku$ overstates flow or understates density. Travel-time and delay studies, including the moving-vehicle method used in Question 6 of this paper, therefore report space mean speed, while spot-speed studies for setting speed limits report time mean speed and its 85th percentile.

Part (c) — Circular versus spiral curves. A circular (simple) horizontal curve is an arc of constant radius $R$ joining two tangents; its curvature ($1/R$) jumps discontinuously from zero on the tangent to $1/R$ at the point of curvature. Because the lateral acceleration required of a vehicle is $v^{2}/R$, that discontinuity demands an instantaneous change in centripetal force, which a driver can only supply by steering across the lane while entering the curve — the well-known involuntary "cutting" of the first part of a curve. A spiral (transition) curve — almost always a clothoid, for which curvature increases linearly with distance along the curve, $1/R = \ell/(A^{2})$ — is inserted between the tangent and the circular arc so that curvature, and therefore lateral acceleration, builds up gradually.

The spiral serves three purposes at once. It provides a natural path that matches what drivers actually steer, so lane encroachment is reduced. It provides the length over which superelevation can be run out and the pavement rotated from normal crown to full superelevation, with the rate of rotation tied to the rate of curvature change rather than applied arbitrarily on the tangent. And it improves appearance, avoiding the kink that a bare tangent-to-arc junction shows in perspective. The cost is geometric complexity: the circular arc must be shifted inward by the shift $p \approx L_s^{2}/(24R)$, the tangent distance lengthens, and staking requires spiral tables or coordinate geometry. Practice in the TAC Geometric Design Guide is to spiral high-speed curves and sharp curves, and to omit the spiral on flat curves at low design speed where the superelevation run-out can be accommodated on the tangent.

Part (d) — HOV lanes. A high-occupancy-vehicle lane is a lane reserved, either full-time or during stated hours, for vehicles carrying at least a specified number of occupants (typically 2+ or 3+), usually together with transit buses, and in many Canadian jurisdictions motorcycles and permitted low-emission vehicles. The lane may be a converted general-purpose lane, an added lane, a barrier-separated facility, a contra-flow lane on the opposing carriageway during a directional peak, or a queue-jump at a ramp or intersection approach.

The rationale is that a freeway lane is capacity-limited in vehicles per hour but the transport task is measured in persons per hour. A general lane carrying 2,000 veh/h at an average occupancy of 1.15 moves about 2,300 persons per hour; an HOV lane carrying only 900 veh/h at an occupancy of 2.5, plus a dozen buses at 50 passengers each, moves well over 2,800. The lane therefore raises person-throughput while carrying fewer vehicles, and the travel-time advantage it offers is the incentive that recruits carpools and transit riders in the first place. The design issues are enforcement (violation rates rise sharply when the lane is not separated or not patrolled), the differential-speed hazard where an uncongested HOV lane runs beside a congested general lane, weaving at the ingress and egress points, and the political difficulty of converting an existing general lane rather than adding one. An HOV lane that is under-used is a genuine capacity loss and the occupancy threshold should then be relaxed; one that is itself congested has succeeded and should be tightened.

Part (e) — Pedestrian clearance time. Pedestrian clearance time is the interval, displayed as a flashing DON'T WALK (flashing orange hand), that begins at the end of the WALK indication and allows a pedestrian who stepped off the kerb on the last instant of WALK to complete the crossing before conflicting vehicles are released. It is computed from the crossing length and an assumed walking speed:

$$\mathrm{PCT}=\frac{L_c}{S_p}$$

where $L_c$ is the crosswalk length measured from kerb to the far kerb (or to a median refuge where one is provided) and $S_p$ is the design walking speed. Canadian and HCM practice has migrated from 1.2 m/s to 1.0 m/s or slower where older pedestrians, schools or hospitals dominate the demand. Pedestrian clearance is only one part of the total pedestrian requirement; the minimum green a phase must display is

$$G_p = t_{\text{start-up}}+\frac{L_c}{S_p}+2.7\,\frac{N_{\text{ped}}}{W_E}$$

with $t_{\text{start-up}} \approx 3.2$ s of perception-reaction and start-up, and the last term accounting for the time a platoon of $N_{\text{ped}}$ pedestrians needs to discharge through a crosswalk of effective width $W_E$. It is essential to distinguish the clearance interval from the WALK interval: a signal whose total pedestrian time is adequate but whose flashing DON'T WALK is short will strand slow pedestrians in the roadway even though the arithmetic sum looks right. Questions 2 and 3 of this paper show the consequence at the network level — on a heavily-walked intersection this relation, not Webster's vehicle optimum, sets the cycle length.

Topic. The vocabulary layer of traffic engineering: signal phase ordering, the two speed averages, horizontal transition geometry, priority lane management, and pedestrian timing.

Key relations. $u_t = u_s + \sigma_s^{2}/u_s$ (time versus space mean speed); $q = k\,u_s$ (fundamental relation, space mean speed only); clothoid $R\ell = A^{2}$ with shift $p \approx L_s^{2}/(24R)$; $\mathrm{PCT} = L_c/S_p$; $G_p = 3.2 + L_c/S_p + 2.7 N_{\text{ped}}/W_E$.

Why this works. Each definition rests on a single physical idea. Leading versus lagging is about when a queue is discharged relative to the conflicting stream, so it is a queue-storage and expectancy question, not a capacity question. The two speed averages differ because sampling at a point weights vehicles by their own speed, which is exactly the bias the harmonic mean removes. A spiral exists because curvature, not radius, is what a driver must change at a finite rate. An HOV lane exists because the conserved quantity of interest is persons, not vehicles. And pedestrian clearance exists because a pedestrian's exposure is a length divided by a speed, whereas a vehicle's is a reaction plus a deceleration.

Common pitfall. Substituting time mean speed into $q = ku$ — it is the single most common error in this subject and it biases every derived density. A close second is treating pedestrian clearance time as the whole pedestrian requirement, ignoring the start-up and platoon-discharge terms, which under-times a busy crossing by ten seconds or more.

Source. Garber & Hoel, Traffic and Highway Engineering, 5th ed., §6.3 (speed studies and the two means), §3.4 (horizontal alignment and spirals), §8.4 (signal phasing); HCM, signalised-intersection pedestrian methodology; TAC Geometric Design Guide for Canadian Roads, Ch. 2 (alignment) and Ch. 4 (HOV facilities).

Question 1 of 7: Definitions and Discussion

Question 1: Definitions and Discussion (5 × 4 = 20 marks)