Peter Freeman wrote: ↑Thu Oct 10, 2019 03:02
This is an interesting and objective report. What it stresses though is that the (shorter journey times and congestion reduction) benefits almost always
do result at peak times, but are countered by dis-benefits at off-peak times. The benefits balance in most cases is positive, but that doesn't make off-peak users happy.
Journey time 'benefits' are on the whole slightly negative across the 54 schemes (monetised as £5.1m gain in peaks, £5.4m loss offpeak).
It is important to note that when this report refers to signalisation, most of its examples are the signalisation of roundabout approaches, not replacement with cross-roads or T's. Replacement introduces one (perhaps long-ish) delay at the signal. Full signalisation of a roundabout introduces multiple stops (depending on which turn you are making, and subject to the quality of signal phase tuning). This, amongst other reasons, is why I am so critical of signalisation of existing roundabouts, while acknowledging that it is a cheap and cost-effective quick-fix.
Yes, I agree signalised intersections are much more efficient at moving traffic than signalised rbts. The problem is that they don't have the safety benefits (lack of deflection etc). So I don't think that's a one size fits all answer - it will be site specific.
Peter Freeman wrote: ↑Thu Oct 10, 2019 03:59
I agree, left-turn filter lanes are, in many cases, the very best quick capacity boost. They are so good that they should almost be regarded as a standard part of initial roundabout design. In Australia (where we do build roundabouts, and where they work perfectly well in the right circumstances, as do yours) many roundabouts have filter lanes from commissioning day. Is the UK also doing this? - most that I know of are retrofits.
Absolutely on board with you. Whereas the choice between rbt, signalised rbt and intersection is quite subtle and context specific IMO, the choice between filter lane and no filter lane is not a hard one for a busy movement. They should be the 'default' improvement, not signals.
As for building them from the start, I'd like to see that much more as well, though it does happen. Here's a nice example in Watford:
https://www.google.co.uk/maps/@51.65521 ... a=!3m1!1e3
The filter lane downside is that it exacerbates what I see as an inherent roundabout problem. Each entry's give-way line used to be (in the idealised world of the 1960's …!) the beginning of a short low-speed weaving section, where vehicles safely juggled each other into the correct positions to either exit or continue to circulate. This mode of operation still occurs in urban gyratories. But nowadays roundabouts (UK) are large, they operate at high speeds, and the entry/exit proximity leaves no weaving possibility. Thus, the give-way line is a 'starting-grid' from which to launch a high acceleration forced entry into an unforgiving traffic platoon. It's actually a glancing-angle cross-roads of two one-way roads. While waiting at that give-way line you're looking for a gap that might be created by someone using the exit that lies just to your right, but often it's impossible to predict the intention of an approaching vehicle until it is too late to exploit the gap. If you plan to turn immediately left you might be tempted to do it in the belief that the approaching vehicle is continuing on the circulatory carriageway, but this is often unclear (again until too late). Also, if you plan to turn immediately left you might be tempted to proceed in the belief that you can perform a safe side-by-side usage of the exit, but will the approaching vehicle actually stay in-lane and allow that to work?
I think I'm saying "Hey, roundabouts don't work so well after all" … ? (At least, big UK ones on major roads). (I drive frequently in the UK).
I see what you're saying about roundabouts generally (except I'm not sure how filters could make things worse?). To be honest, the roundabouts I know that have been signalised were quite intimidating to enter pre-signalisation. That said, I never actually saw a collision, and if there were any they would have been low speed side-swipe type affairs. More than anything I marvelled at how well they flowed given what they had to deal with.
It is absolutely true that the signalised replacements feel safer, but in the case of the crossroads at least I suspect that when there are collisions they may be more serious. I've seen cars miss red lights at signalised crossroads, and been driven through a red at speed once overseas, and though this is a low frequency event it poses a completely different level of danger to anything at a roundabout.
I saw a study along these lines once upon a time (sadly don't recall where) which basically said that signalised crossroads are the best of both worlds from a safety perspective, which is plausible but unfortunate given their negative journey time effects.