Question 5 of 5: Advanced Public Transportation Systems and Transit Priority

Question 5: Advanced Public Transportation Systems and Transit Priority (20 points)

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.

(a) APTS components, data needs and acquisition techniques (12 points)

Advanced Public Transportation Systems apply the same sensing, communication and control technologies to the transit fleet and its passengers that ATMS applies to the road network. The organising insight is that a transit system is a scheduled service whose quality is judged by reliability rather than by speed, and that reliability cannot be managed without knowing, continuously, where every vehicle is relative to where it should be. Every APTS component either produces that knowledge or acts on it.

Components. On the vehicle, the core is a GNSS receiver augmented by dead reckoning from the odometer and a rate gyroscope, so that position survives the urban canyons and tunnels where satellite geometry fails; an on-board computer and mobile data terminal that report position and receive dispatch messages; automatic passenger counters at each door; a fare validator for contactless smart cards or account-based media; and a security and maintenance layer comprising on-board cameras, a silent alarm, and engine and subsystem health monitoring drawn from the vehicle data bus. At the centre, the automatic vehicle location function displays the fleet against the schedule, and computer-aided dispatch gives the controller the means to act on what it shows — holding a vehicle, short-turning it, inserting a relief vehicle, or dispatching maintenance. Around them sit the scheduling and run-cutting system that produces the timetable, the maintenance management system, and the prediction engine that converts vehicle positions into arrival times. On the wayside and in the traveller's hand are real-time arrival displays at stops and stations, the public feed and mobile applications, fare gates and off-board payment machines, roadside units and emitters that receive signal-priority requests, and park-and-ride occupancy detection.

Data needs. The primary datum is vehicle position with an accurate timestamp, reported often enough that a prediction can be made between reports — a few seconds on a frequent urban route. From it are derived schedule adherence, headway between successive vehicles on the same route, and running time by segment. Dwell time at each stop is needed separately, because dwell is both a large share of running time and its most variable component. Boardings and alightings by stop and by time of day are needed for planning and for load management; passenger load is needed in real time if priority is to be granted conditionally on occupancy. Fare transactions are needed for revenue and, where the system records both boarding and alighting, they yield an origin–destination matrix of a quality no survey can match. Priority request and grant logs are needed to demonstrate that the signal priority investment is working. Vehicle health and fault codes are needed to move maintenance from a fixed interval to a condition basis.

Acquisition techniques. Position is acquired by GNSS with dead-reckoning augmentation, and legacy fleets may still use signpost-and-odometer systems in which a wayside beacon resets the odometer count at known points. The uplink is by cellular data or a private mobile radio network, with the trade-off being recurring cost against capital cost and coverage control. Passenger counting is by infrared beam pairs across the door, by treadle mats in the stepwell, or — now the preferred technique — by overhead stereo-vision or time-of-flight cameras, which count in both directions and tolerate crowding far better. Fare data is acquired by contactless smart-card or bank-card readers, and where the fare structure requires a tap on exit the same readers yield the origin–destination data. Vehicle health is acquired from the vehicle data bus. Priority requests are issued by short-range roadside communication, an optical emitter, or directly from the centre over the network. Origin–destination and transfer behaviour can be supplemented by Wi-Fi or Bluetooth address re-identification on board, subject to the same privacy design as its roadside counterpart.

(b) Transit priority measures and the techniques that implement them (8 points)

Transit priority is the deliberate reallocation of road space or signal time from general traffic to transit vehicles, justified by the fact that a bus carrying fifty passengers has fifty times the person-throughput of a car carrying one, so that person-delay falls even when vehicle-delay rises. The available measures form a progression: passive measures cost only staff time, active measures require detection and a capable controller, and physical measures require road space and a political decision.

Passive priority changes the signal timing permanently in the transit vehicle's favour without detecting it at all. It includes setting the progression speed of the coordinated system to the bus operating speed rather than the car operating speed; shortening the cycle length on the transit corridor, since shorter cycles reduce average delay to any vehicle that must wait; relocating near-side stops to the far side of the intersection, so that a bus is not obliged to serve a stop and then wait through a red it could have cleared; and consolidating closely spaced stops, which removes dwell events entirely. These are the cheapest measures available and are routinely overlooked.

Active priority detects the transit vehicle and modifies the current cycle for it. The two basic strategies, shown below, are green extension, in which a bus approaching as the green is about to end is given a few additional seconds so that it clears without stopping, and early green or red truncation, in which a bus waiting on red has the conflicting phase cut short so that its green returns sooner. Beyond these, a transit phase may be inserted into the cycle or the phase order rotated so that the movement the bus needs is served earlier.

The technique that implements active priority is standardised. A priority request generator on the vehicle, knowing its own location, route and schedule adherence, issues a request; a priority request server in the signal controller receives it, arbitrates between competing requests and against emergency pre-emption, which always outranks priority, and grants or refuses it. NTCIP 1211 defines the objects and the message exchange, and specifying conformance to it is what allows buses and controllers to be procured separately. The request can travel by short-range roadside communication, by an optical emitter of the kind long used for emergency pre-emption, by a loop or radio-frequency tag, or over the network from the transit centre to the traffic centre, which is the architecturally cleanest path because it lets the traffic agency see and log every request. Two refinements decide whether priority is politically sustainable. The first is conditional priority: the request is issued only when the vehicle is behind schedule by more than a threshold, or is carrying more than a threshold load, so that priority is spent where it buys reliability rather than on every vehicle. The second is cycle recovery: the controller repays the borrowed time over the next two or three cycles so that coordination for the cross street is restored, which is what prevents priority from degrading the rest of the network.

Physical and operational measures deliver the largest and most reliable gains. Dedicated bus lanes, whether curb-running, median-running or contraflow, remove the bus from the general-traffic queue altogether. A queue-jump lane combined with a leading bus signal — a short transit-only interval that releases the bus a few seconds ahead of the parallel traffic — achieves much of the benefit of a full bus lane at one intersection's cost. Bus bulbs and boarding islands remove the delay of pulling out of and back into traffic. All-door boarding with off-board or on-board tap fare payment, level boarding and low-floor vehicles attack dwell time, which on a busy urban route is both the largest controllable component of running time and the most variable, so that reducing it improves reliability more than any signal measure. Operationally, headway-based holding control — instructing a vehicle that is running close behind its leader to wait — is the standard remedy for bus bunching, and requires only the automatic vehicle location data that part (a) described.

In Canadian practice these measures are usually deployed as a package rather than singly: a rapid-transit corridor typically combines a dedicated lane, far-side stops with bulbs, all-door boarding, conditional signal priority at every intersection and real-time arrival displays, and reports its benefit as a reduction in the ninety-fifth-percentile running time rather than in the mean, because reliability is what the riders notice.