MOVA at stand-alone crossings
Author: Mark Crabtree (TRL Limited)
Tel: +44 (0)1344 770959
Last Updated: 09/03/2006
MOVA is highly suitable as a control strategy for all types of signal-controlled stand-alone crossings for pedestrians, cyclists or horses. MOVA can be used at any type of crossing (ie Pelican, Puffin, Toucan, Pegasus or even the unofficial ‘Onecans’).
At sites falling into the high speed category (ie 85th percentile approach speed in excess of 35mph) full MOVA has to be used (ie including IN-detectors). Such sites would require full D-system VA with SA/SDE if MOVA was not fitted. At other sites, Compact MOVA may be used (ie just using X-detectors without IN-detectors) – this requires MOVA M5 to be used.
Usually, setting up a stand-alone crossing in MOVA is relatively straightforward compared to setting up a junction. The most important aspect to get right is the cruise speed. There will be a very high proportion of optimised changes and changing at a suitable time in the face of approaching traffic is crucial to effective and safe operation. Accurate saturation flow values will also help (but no more so than at junctions). Be prepared to use different values by time of day especially at sites in urban areas where late night revellers may be combined with slightly higher speeds for example.
To operate a crossing in a conventional way, that is to simply direct pedestrian demands into MOVA, it is normal to get the signal controller do the work of providing a wait-light mimic. The location of the wait-light mimic is specified in the LTYPE(link) array. If the demand is to be latched (ie where there are no kerbside detectors and therefore no opportunity to cancel demands) the LTYPE(link) is simply the detector channel number of the pedestrian demand. In MOVA M5 and later there is a complementary array called WAITCH(link). This should be left as zero in the above case.
If kerbside detectors are present, the controller needs to generate and cancel the demand as appropriate, and pass it to MOVA. The value in LTYPE on this occasion needs to be the detector channel, plus 100, to specify the unlatched situation which allows MOVA to cancel the demand. Again the WAITCH (MOVA M5) can be zero if the wait-light mimic is specified in LTYPE. SDCODEs for any such links must be unlatched ones, usually 3 for two-stage crossings. If using MOVA M4, channel 24 also needs to be specified in LTYPE and be either set permanently ‘on’, or also to the wait light mimic. LTYPE can be either ‘24’ or ‘124’ and the SDCODEs for this link must all be zero.
Alternatively, in MOVA M5, it is possible to pass the wait light mimic from the controller to MOVA, and separately connect the kerbside detectors. In this case, each kerbside detector needs its own channel, which needs to be specified in LTYPE(links) as the channel plus 100. The WAITCH for the links needs to be the wait-light mimic detector channel (NOT plus 100). Both the kerbside detector and the wait light mimic then have to be ‘on’ for a demand to be generated for the link. The reason for configuring a crossing in this way is described in the next paragraph.
For links with pedestrian detection (whether latched or unlatched) the STOPEN value can be set to 64 or 65. This has the effect of adding a time-weighting into MOVA’s stop and delay optimisation process, with ‘65’ ramping-up faster than ‘64’. Hence when MOVA is optimising, it is taking some notice of the fact that pedestrians are waiting, trying progressively harder to find a gap as time elapses. In the case where the kerbside detectors are separately connected, the effect on the stop and delay optimisation can then depend on the number of detectors that are currently detecting – in effect this provides a crude representation of the number of pedestrians waiting. Hence MOVA will try harder to find a gap if there are two detectors on, rather than one. The feature can work with MOVA M4 – in which case the kerbsides are connected directly, channel 24 has to be the wait-light mimic and has to be LTYPE = 124. Note that in MOVA M4 and M5 this feature only really works with two-stage situations. In MOVA M6 the feature will work at junctions as well.
Where a crossing is split into two with a central island with two separate phases, a single MOVA stream can be used which will need to have three or four stages with suitable demands being set. However, with dual stream MOVA becoming available without extra cost, it is possible to control the two parts completely separately if desired.
In general, the behaviour of MOVA at stand-alone crossings is safe and very efficient in comparison to VA, fixed time or possibly SCOOT even. VA tends to need quite large gaps in which to change stage, and the gaps are often big enough to tempt pedestrians to cross before they get a green man. With SA/SDE included, the gaps needed to change are often easily big enough for most pedestrians to cross quite safely.
UTC systems usually run cycle-times that are efficient for junctions, hence rather long, leading to long waiting times for pedestrians. There are ways to alleviate this somewhat, by allowing a longer or multiple windows in which the stage can run. However, this will have the problems that pre-timed maximum VA operation has (see below). Where VA crossings are within a UTC system, there is a tendency for the signals to change in a big gap, just ahead of a major platoon. Obviously this is not ideal!
In contrast, MOVA does not wait for an excessively long gap. It will normally choose a gap that allows traffic to stop easily enough, but not large enough to tempt a pedestrian into crossing. The change will often assume that one or more vehicles will continue when leaving amber is presented. This greatly helps in allowing an appropriate gap to be found and makes operation more responsive to the pedestrian. Such positive control helps avoid confusion to both traffic and pedestrians which is normally beneficial to safety. Within a UTC system, the quicker reaction to gaps should reduce the problem of changing inappropriately in the face of a major platoon; on the other hand, an approaching platoon will be detected sooner (in standard MOVA) when they cross the more distant IN-detectors.
Pre-timed maximum VA control is frequently used at stand-alone crossings. On the face of it this is an efficient form of control for pedestrians as they will often get instant service at crossings that are not too busy. However, at very busy crossings where pedestrian demands are frequent, VA control still suffers from all of the problems outlined above, leaving many pedestrians to wait until the maximum has timed-off. At quieter sites, the instant change when a demand is made means that leaving amber may be presented at a time when drivers may not be able to stop comfortably. MOVA (including Compact MOVA) should provide a safer alternative by choosing a gap in the traffic, thereby increasing the chances of drivers obeying the signals.
Compact MOVA should be extremely effective at urban crossings, being particularly responsive to pedestrian demands, but not penalising traffic in comparison with VA. Such rapid response even works well at busy sites in comparison to pre-timed max control. It also reduces reliance on kerbside detection and call-cancelling as any pedestrian genuinely wanting to cross the road is unlikely to be able to do so until the signals have changed! There is a need to place the X-detectors a bit further upstream than Standard MOVA to help ensure vehicles on a critical part of the approach are dealt with appropriately. For more information see AG45 issue C.