A good question from the crowd is always worth investigation and one that didn’t have an answer immediately to hand was: how much is traffic slowed by a pedestrian crossing? So here is a quick set of numbers for an example crossing comparing the typical treatments to each other.
Taking the existing tuned traffic model it was applied to a loop of road 400m in length having a speed limit of 60km/h (30 mph) with a pedestrian zebra crossing, the pedestrians are a simpler model than the traffic and assume no constraints to their throughput from inadequate footpath widths etc and take their distribution of speeds from data measured in Australia. Checking for the onset of congestion at different rates of pedestrians yields this:
At the lowest rate averaging a pedestrian every 2 minutes already the peak throughput is obviously reduced, increasing to 300 per hour or an average of 1 pedestrian every 12 seconds the throughput is less than half the unimpeded peak. Cars stopping, starting, and forming queues are very inefficient so the small number of pedestrians crossing with priority quickly chokes the traffic flows. Of course if anything gets in the way of traffic throughput it must be fixed and from the Victorian guides:
[zebra] crossings give pedestrians priority over vehicles. Delays to both pedestrians and vehicles may be minimised compared with pedestrian operated signals, although vehicle delays can increase considerably when pedestrian flows become very high.
Not the easiest to understand phrasing and I’m not sure I’d agree with their consideration of a pedestrian every 12 seconds to be “very high” but lets now look at the case of a signalised crossing. The geometry of the crossing remains unchanged as a 4m wide footpath continuing across 2 lanes of traffic in a single span of 6m, but the principal parameters determining the performance is the timing of the phases. Taking an example from Melbourne the pedestrian cycle is 20 seconds split between 10 seconds of walk signal and 10 seconds of clearance time, the timeout before another pedestrian phase is 60 seconds, and the inter green period to stop traffic is 4 seconds. This is a total cycle time of 84 seconds, with 60 of that clear for motor traffic roughly matching the 70% throughput achieved. Plotting the peak throughputs (achievable transport utility) as a fraction of 2200 vehicles/lane/hour we can see these pedestrian crossing lights approaching that simple estimate.
Many jurisdictions are now installing various types of “intelligent” pedestrian crossing lights with various poorly defined soft names to appeal to the public, but they’re once again advantages solely for motorised traffic dressed up with ridiculous claims. Victorian guides say:
the clearance phase is shortened when the crossing detectors sense that a pedestrian has crossed quickly, reducing delays to vehicles, the clearance phase is lengthened when the crossing detectors sense pedestrians still on the crossing towards the end of the normal clearance time. This is an advantage for slower pedestrians
Except they are already required to design for the slowest users, and (although they refuse to) safety for all users is the basic requirement for traffic management and saying an intelligent crossing is required to allow this safety is shirking their responsibility with excuses to create more work for themselves or as previously disregard public safety in the interests of motor vehicle throughput.
Still reinforcing the blind use of technology is a claim that:
Following conversion of a [signalised crossing] to [an intelligent] Crossing in an eastern suburb of Melbourne, driver delay was found to have been reduced by around 50%.
When the conditions to reduce delay by these impressive amounts would indicate the volumes of both traffic flows are so low that a zebra crossing would reduce delay even further, most substantially for pedestrians when the priority returns to them. The intelligent crossings do improve motorised traffic throughput and reduce motorised traffic delays but they add no improvements for pedestrians. Examples that could be added for pedestrians benefit might include:
- A new phase where the red pedestrian crossing signal is extinguished (show no pedestrian signals) and pedestrians are free to cross at will, giving way to traffic.
- Sensing approaching traffic and releasing waiting pedestrians to cross with a green signal while there is no traffic.
Both remove the current situation in Melbourne of pedestrians patiently waiting at crossing lights while there is no traffic moving past, only to be given their green phase and stop approaching motor vehicles unnecessarily. The intelligent crossings can be programmed to cancel the pedestrian request if they sense the pedestrians have crossed against the red light, but that relies on pedestrians breaking a law which gets frequent police “focus”. Instead we should be following the lead from other countries in minimising delay for all users, not just focusing on motorised transport. Plotted as the example +Fast Start, its possible to include these advanced pedestrian movements without reducing motor vehicle throughput by managing transfer of share to/from following stages (already implemented in some NSW locations), these techniques are proven even in coordinated/sequenced/managed systems.
The control by simple duration timers and focus solely on reducing motor vehicle delay at Melbournes crossings is reinforcing private car use and demand for parking immediately adjacent to the travel destination, putting more cars on the road. Traffic congestion is not solvable by further cementing the dominance of cars and increasing their throughput, but by prioritising and making other modes that use less space more attractive.