Three levels of bus priority are provided in SCOOT to benefit a bus, or other "priority" vehicle:
- Equal priority for all buses on an approach by extending the current stage to allow it clear the junction, or shortening intervening stages to return more quickly to the bus stage.
- Differential priority: extra priority for some, less (or none) for others. E.g. no priority for buses running on time, moderate priority for late buses, high priority for very late buses. The bus must be able to indicate its category to SCOOT.
- Extra priority by skipping non-bus stages (MC3 only).
Detection and identification
The logic of bus priority does not depend on the method of detecting buses. SCOOT can use detection from all sorts of bus detection systems: the increasingly popular automatic vehicle location (AVL) used by many bus management systems, selective vehicle detectors (SVD) or any other system that can reliably detect a bus at a known position. Where SCOOT is given a bus identifier as part of the bus detection, it can match this detection with a previous detection of the same vehicle.
If differential priority is required, then the bus detection system must provide the required priority level as well as the detection of the bus. For a typical system that adjusts the priority in accordance with the lateness, or adherence to schedule of the bus, the bus management system will define the lateness and send it to the bus. When the on-bus AVL system determines that the bus is at the bus priority detection point, it will transmit a bus detection, priority level and possibly a vehicle identifier. SCOOT will then provide the appropriate priority as defined by the traffic engineer in the SCOOT data.
Buses normally need to be detected after any bus stops on the link as SCOOT does not attempt to model the time spent at a bus stop. Accurate location, to detect the bus as it leaves the bus stop can be critical in this area. Bus loops or AVL systems where the bus detection points can be specified are best as they can be set to detect in the optimum position. Where the time spent at the bus stop is predictable, e.g. buses with many doors, pre-paid tickets and all buses expected to stop, it may be possible to detect before the stop and allow for the time spent at the stop.
Buses are modelled according to the road layout. Where there is no bus lane, they are modelled as queuing with other vehicles, allowing them to be given priority even though they may be delayed by other vehicles. Where there is a bus lane, buses are modelled separately from other vehicles, but will be modelled to join the general traffic at the end of the bus lane if it ends before the stopline.
The signal timings are optimised to benefit the buses by either extending a current green signal (an extension) or causing succeeding stages to occur early (a recall). In SCOOT MC3, recalls can be enhanced by stage skipping. That is, omitting stages between the current stage (shortened by the recall logic) and the bus stage.
Extensions can be awarded centrally, or the signal controller can be programmed to implement extensions locally on street (a local extension). SCOOT can be configured by node to allow or disallow each of these methods of priority. In principle recalls could also be awarded locally, but the timing is less critical and the extra programming of the controllers is not considered cost effective.
Extensions awarded in the controller can be advantageous as they eliminate 3 to 4 seconds transmission delay from street to computer and back to street, and so allow the system to grant extensions to buses which arrive in the last few seconds of green. This is especially important in London where link lengths are short, with bus stops often further restricting the effective length. SCOOT is still in control as it sends a bit each second to permit local extensions only when the saturation of the junction is sufficiently low. Techniques for programming the signal controller have been developed and implemented in London.
Once the bus has passed through the signals, a period of recovery occurs to bring the timings back into line with the normal SCOOT optimisation. Four methods of recovery are provided for operation after extensions, recalls and skips, of which two methods (one for extensions and one for recalls and skips) are recommended for normal use and operate by default.
Restrictions on priority
The amount of priority given to buses can be restricted depending on the saturation of the junction as modelled by SCOOT. This is controlled by target degrees of saturation for extensions and recalls. These are the degrees of saturation to which the non-priority stages can be run in the case of a priority extension or recall respectively. For stage skipping, the limit is set on the current degree of saturation of the skipped stage. Normally the target saturations should be set so that the junction is not allowed to become oversaturated, although some degree of oversaturation may be allowed to service an extension, or a stage skip where the skipped stage is not busy. This means that bus priority will be most effective at junctions which have spare capacity.
Congestion Management Using Gating for bus priority
The Bus SCOOT applications described in the previous sections aim to reduce "phase delays" (due to buses arriving at a junction on red), rather than the more significant delays which can occur to buses queuing to reach the junction in congested conditions. Congestion management strategies for buses, using the SCOOT "gating" facility, have therefore been developed and evaluated.
Gating (also known as traffic metering) is a traffic management technique which allows queues to be relocated away from one or more congested links within a network. For bus priority, the queues would be relocated onto one or more upstream links where it is more feasible to protect buses by physical bus priority, such as a bus lane. Typically, gating is used to hold traffic outside a town centre to maintain free movement of vehicles in the central area. It is hoped that, in keeping internal, critical, links relatively free of congestion, the network becomes more stable with positive effects on public transport:
- bus journey times become more reliable
- buses will be able to enter links more easily
- buses will be able to pull out from bus stops more easily
- delay is reduced for buses
SCOOT gating is used to control the rate at which traffic is allowed into the previously congested area. Buses are protected from the relocated upstream queue by the bus lane. When the correct metering balance is achieved, transit times for buses through the network reduce, without imposing increased transit times on general traffic.
Trials of SCOOT bus priority in conjunction with congestion management facilities (SCOOT gating, bus lanes) have been carried out in both the Twickenham and Edgware Road in London. The Twickenham trial demonstrated the benefits of using gating in conjunction with bus priority. The trial at Edgware Road was less successful in that the gating strategy increased the overall delays. However the work did demonstrate the advantages of combining gating with bus priority. The following conclusions were drawn:
- Network characteristics: gating is most beneficial to general traffic where there is a substantial amount of cross-movement traffic flow, e.g. where north-south traffic conflicts with east-west traffic. Conversely, gating is less effective on arterial roads where the large majority of traffic is travelling in the same direction.
- Bus lanes: public transport gains most when the gated link(s) has/have a bus lane, allowing buses to bypass queues.
- Benefits will be maximised by restricting priority to late/long headway buses.
To view the results of surveys of SCOOT bus priority click here