The hi-viz SMIDSY thread

[Hot_Air]

This thread has been conspicuous by its absence (until now).

Here’s a thing about science: people change their mind as new evidence emerges. Now I’ve read the Science of Being Seen, and – because I decided hi-viz didn’t work – I wear black leathers.

Maybe I made the wrong choice after all. Horse shared a systematic review by TRL, which he co-authored; I think the conclusion inferred that further research on hi-viz wasn’t warranted because there’s sufficient evidence of benefit. (Yes, I know there are caveats about this, and it’s nuanced.) I'm sure Horse will correct me

Then there’s saccadic masking. It transpires that saccades don’t hamper car drivers from seeing bikers: research on visual search has shown this. (It’s peer-reviewed science, but I’ve lost the reference, unfortunately.)

Enter research about cyclists. I know, they wear lycra and don’t go ‘Vroom!’ But there’s a randomised controlled trial (RCT) about hi-viz. In scienceland, an RCT is a type of study that’s a Big Deal. It’s the most rigorous type of trial, considered the highest level of evidence (excepting a meta-analysis of several RCTs). And this RCT investigated whether a hi-viz jacket affected the number of accidents, not merely the effect on 'detectability'. So, even if I don’t want to like the findings, I think it’s worth reading with an open mind: The effect of a yellow bicycle jacket on cyclist accidents: a randomised controlled trial

[Horse]

Always a fun and factoid-filled topic for clearing a forum

My views? I started riding when bike struggled to run a 35W (sometimes even 30W) dipped beam at night, I was probably an early adopter - to the extent that drivers often (ok, I exaggerate for comic effect, it was a few times) wound (good old twirly handle) their window down to point out "Ear, mate, yer edlampson!" and look confused when told that yes, it was supposed to be.

It had shock value then. Now lost.

In about 1992 I bought a white KRT, at the time when plod still had plain white bike. Mine had no drl, no stripes, etc, I didn't wear hi-viz. Many (oh alright, a few) times, drivers pulled out in front of me, then you could see the muscles in the back of their neck tense as a little part of their brain went 'tap tap err hello, do you realise what you've just done?' Other times, they would look, start to pull out, think, then stop half way.

I posted on Vd, years ago, two stories from trafpols, 'Danno' (from "book em") and (now the late) Martin.

Danno was riding along in a 30 in basingstoke, minding his own business. Car pulled out close in front, very close. They stopped for a chat.

Martin was filtering on a dual, both lanes stopped. Driver ahead in left lane looked in the door mirror, then edged his car out to block the gap. When traffic got moving, they stopped for a chat.

In both cases, the drivers said exactly the same:

"Sorry, officer, I thought you were an ordinary rider".

Aids might work, but you'd but daft to rely on it.

[Horse]

Possibly supported by work the Walker chap at Bath did, dressing cyclist in different outfits. IIRC drivers gave the the 'police' jacket rider the widest passing room, but closest to lycra-clad riders.

Edit:
This study looked at whether drivers overtaking a bicyclist changed the proximities of their passes in response to the level of experience and skill signalled by the bicyclist's appearance. Five outfits were tested, ranging from a stereotypical sport rider's outfit, portraying high experience and skill, to a vest with ‘novice cyclist’ printed on the back, portraying low experience. A high-visibility bicycling jacket was also used, as were two commercially available safety vests, one featuring a prominent mention of the word ‘police’ and a warning that the rider was video-recording their journey, and one modelled after a police officer's jacket but with a letter changed so it read ‘POLITE’. An ultrasonic distance sensor recorded the space left by vehicles passing the bicyclist on a regular commuting route. 5690 data points fulfilled the criteria for the study and were included in the analyses. The only outfit associated with a significant change in mean passing proximities was the police/video-recording jacket. Contrary to predictions, drivers treated the sports outfit and the ‘novice cyclist’ outfit equivalently, suggesting they do not adjust overtaking proximity as a function of a rider's perceived experience. Notably, whilst some outfits seemed to discourage motorists from passing within 1 metre of the rider, approximately 1-2% of overtakes came within 50 cm no matter what outfit was worn. This suggests there is little riders can do, by altering their appearance, to prevent the very closest overtakes; it is suggested that infrastructural, educational or legal measures are more promising for preventing drivers from passing extremely close to bicyclists.

[Horse]

The test group had 47% fewer multiparty accidents with personal injury.

From your link. They still had crashes - did the type etc. vary and level of injury?

[hilldweller]

I firmly believe the new DRL are a killer. Why ?

I was returning from the NEC Bike Show a few years ago. Riding a white ER6n which of course has permanent dipped beam. It was dusk. I turned off the M6 at Sandbach heading for Congleton. I got as far as the other side of the motorway when BANG and I was off. A Porsche 911 had turned off the motorway into my patch and I hit his drivers door.

I was wearing a Frank Thomas bright yellow thermal oversuit, it was November !

I remember standing there weighing up the situation and looking at the traffic I realised I would have been virtually invisible because of the huge glare from cars following me with dipped headlights and the DRL creating a wall of white.

Bike was a write off, 3 weeks later all I was riding a VStrom 650.

The other down side of DRL in daylight, they grab the eye, which is what they are supposed to do but that drags the eye from the poor cyclist or pedestrian.

Since then, the two bikes I've owned have been equipped with DRL and I wear a jacket with high viz panels.

A bit of good came from my crash, they reworked the motorway junction with traffic lights so hopefully no more accidents there.

[The Spin Doctor]

Maybe I made the wrong choice after all. Horse shared a systematic review by TRL, which he co-authored; I think the conclusion inferred that further research on hi-viz wasn’t warranted because there’s sufficient evidence of benefit.

REALLY?

The latest studies I've seen suggest the opposite - that there is little reliable work that provides evidence of benefits.

As I have said many times, if it worked, we'd have seen a reduction in junction collisions. It's not happened.

Then there’s saccadic masking. It transpires that saccades don’t hamper car drivers from seeing bikers: research on visual search has shown this. (It’s peer-reviewed science, but I’ve lost the reference, unfortunately.)

REALLY? Is that just ONE study? It's going against a mass of other work if it is, as you'll know if you have waded through the references on SOBS.

Re the randomised study. Don't confuse randomised with blind. All that it did was split a group in two, got half to wear hi-vis and the other half to ride wearing 'regular' clothing.

By the very fact that you're asking someone to wear hi-vis and participate in a trial you've guaranteed that you've just made them more aware of, and almost certainly more alert to, the looked but failed to see issue.

[Horse]

The TRL literature review says:



The majority of early evidence (mainly from the 1970s and 1980s) concerned bright clothing and daytime running lights on motorcycles. When considering the weight of evidence, both seem to be capable of improving conspicuity, when this is measured in terms of detection (under search and attention conspicuity conditions), and when measured in terms of a behavioural response (such as size of gap accepted in front of a given motorcycle). The majority of studies covered in this review support this conclusion, although there are limitations as to how effective any individual interventions can be due to the number of different visual contexts in which motorcyclists find themselves when riding. For example, coloured clothing is more effective when viewed against a contrasting background. In terms of lighting, although it appears that dedicated daytime lighting on motorcycles is effective in increasing conspicuity, this effect may be smaller when other vehicles have their lights on. More research may be needed on this specific issue, however, especially in terms of understanding its impact
on other accident types.

When lighting is arranged in such a way as to accentuate the form of the motorcycle (and to provide greateri nformation for judging approach speed), this aids the observer in determining the time to arrival of the approaching bike (especially at night). However, little real-world research has been done on this specific type of intervention.

Across all treatments there is evidence that colour can play a role in effectiveness; this may be especially true in settings where coloured motorcycle lights aid in the motorcycle standing out from surrounding vehicles which have white lights.

Although most studies reviewed show benefits of bright clothing, dark clothing may be better if the background is also brightly coloured. In line with the underlying mechanisms proposed, higher contrast with background surroundings to enable better visibility, search conspicuity, and attention conspicuity would be beneficial. Given that environments may differ over even fairly small changes in time or location, there is not likely to be a one-size-fits-all solution, meaning that motorcyclists need to be aware of the limitations of whichever interventions they use.

[Hot_Air]

Spin, I think those pesky Canadians have plagiarised your work: Invisibility training for motorcylists

you'll know if you have waded through the references on SOBS.

Yes, I've bought and read the Science of Being Seen. It's an excellent book, and I've recommended it to friends. I'm afraid I have to disagree with one (small) part of it. Regarding saccadic masking, I believe the evidence-base has evolved.

Saccades take milliseconds and only fill in gaps with information that's already present. The gaps are only milliseconds, and there are not enough of them to miss an object entirely. While I no longer have the publications to hand, the relevant research was undertaken in 'visual search'; it's worth tracking down should you decide to update SOBS.

[Horse]

The bigger the visual imprint, the more something stands out and makes an impression on the driver's brain.

Is it really that simple, bigger = more?

For example, how about looking at a sofa and seeing an HGV?
https://www.dropbox.com/s/euud3txjhsvs1 ... v.tif?dl=0

I'm not sure whether the image is publicly available (I saw it in a report that may not have been published), of a car which SMIDSY'd the rear end of a massive, bright red, combine harvester.

Where size does matter is looming. I don't know the eye physiology / biomechanics, whatever, involved, but a larger object will 'pop out' (stop sniggering) sooner, allowing earlier estimation of time to contact. Have a dig around for, I think, Wells, Florida police lighting, it may be in there.

[Horse]

Dark clothes...can create a less distinct profile.

Sorry, mate, that's absolute tosh. To create the 'best' distinct profile (presumably you mean shape or silhouette?) needs two things:
- bike and rider in, as far as possible, completely the same colour
- that colour to be a total contrast to the background

[Hot_Air]

Fair point!

[Horse]

... And it has been suggested that a distinct, consistent, pattern of lights could be used to identify bikes.


https://cordis.europa.eu/project/id/218703/reporting/it


Final Report Summary - 2-BE-SAFE (2-WHEELER BEHAVIOUR AND SAFETY)
Executive Summary:
Up to 2008, the research relating to PTWs’ safety focussed mainly on sociological aspects, training, accidentology and primary safety systems design. Based on the fact that it is crucial to base countermeasures on scientific evidences relating on riders' behaviours and practices, the purpose of 2-BE-SAFE was to contribute to filling the knowledge gap on PTW riders' behaviour. The 2-BE-SAFE project, started on January, 1st 2009, is a focused 36 months research collaborative project, co-funded by the EC, Theme 7 – SST. It involved 27 partners from 11 countries, and was organised in 7 research work packages.
The aim of Work Package 1 was to identify the factors that contribute to PTW crashes across Europe focussing on road infrastructure and weather conditions. 20 prevailing scenarios in 5 European countries that account for most fatal PTW accidents, and the causal factors contributing to crashes have been identified.
Innovative tools have been designed within Work Package 4 : 6 instrumented PTWs, an instrumented car, 2 riding simulators, a driving simulator, and a video-based tool for investigating motorcyclists’ risk awareness.
A pilot naturalistic riding study was conducted within Work Package 2, using the instrumented motorcycles from partners. This naturalistic riding study was a première at pan-European level. The studies successfully tested the proposed naturalistic riding methodology, and developed new knowledge and recommendations for future larger studies. Different approaches to perform automatic event detection of safety critical events were developed, applied and evaluated.
The study implemented in Work Package 3 was focused on PTW riders’ hazard perception and acceptance of risk, and how these characteristics influence their acceptance of new technologies designed to enhance their safety. Significant differences were observed between commuters and sport-riders: commuters tend for example to underestimate the criticality of riding situations. The result of the online survey revealed two groups based on overall acceptance of assistive systems. Levels of acceptance are much lower than for equivalent systems in passenger cars, the study suggests that riders will accept systems that they perceive to be useful, reliable and effective.
Work Package 5 aimed at a better understanding of rider’s behaviours and of critical factors influencing PTW safety. Several experiments focussed on PTWs conspicuity. Results indicated, for instance, that varying riders’ clothing (bright clothes, reflective warning vests, and dark clothes) can enhance riders’ conspicuity in certain situations but the effects are strongly mediated by the background conditions (e.g. lighting conditions) and by the characteristics of the driving situation (e.g. urban vs. rural traffic environment). Variations of specific frontal light configurations were found as promising solutions to enhance PTWs conspicuity. It is proposed to provide a unique visual signature/signal pattern for PTW to other road users Results revealed advantages in terms of a better detection and faster identification for yellow coloured headlights, ABLS (‘Alternating Blinking Light System’) and additional lights on the fork and handlebars for motorcycles (T Light configuration).

[Horse]

Fair point!

Another point is to consider what you hope to achieve? Let's take 'preventing SMIDSYs? If so, then the side on profile / silhouette isn't what a driver needs to be looking for. That's where sleeveless hi-viz waistcoats lose out, because (from the front) not having shoulders and arms is poor for potential conspicuity as you lose the solid outline (contrast with the background yada yada), also adding width may help with looming, the usual grey retroreflective stripe over the shoulders will obliterate any fluoro, plus (ie another minus) the sleeves, fluoro, grey stripes, etc., all are likely to cause edge and surface disruption - camouflage techniques. And then riders often have a fairing or screen so the riders fluoro-clad torso can't be seen anyway.

[The Spin Doctor]

Spin, I think those pesky Canadians have plagiarised your work: Invisibility training for motorcylists

you'll know if you have waded through the references on SOBS.

Yes, I've bought and read the Science of Being Seen. It's an excellent book, and I've recommended it to friends. I'm afraid I have to disagree with one (small) part of it. Regarding saccadic masking, I believe the evidence-base has evolved.

Saccades take milliseconds and only fill in gaps with information that's already present. The gaps are only milliseconds, and there are not enough of them to miss an object entirely. While I no longer have the publications to hand, the relevant research was undertaken in 'visual search'; it's worth tracking down should you decide to update SOBS.

I think you're misunderstanding the issue.

The saccades are the movements between fixations, which are the short pauses where we process the small area at the focus of our foveal vision.

But it seems there's no smooth left-right-left scan of the sort that we're convinced that we use, as far as I can determine from the trials using eye tracking.

Instead my understanding is that we rely on two strategies:

- the first is a learned series of fixations that searches in specific parts of the visual field* via the high resolution foveal vision for specific cues related to the driving activity
- the second are a series of 'on the fly' checks of specific 'objects of interest' detected against the background in peripheral vision which are made via a fixation to examine the object in the high resolution foveal vision

* as I've said elsewhere, the big danger here is that with just a few hours of driving experience we learn that a successful strategy at a busy junction is to search for GAPS, not vehicles. So what we tend to do is to look to the distance where we hope to spot a gap.

So how long do these fixations and saccades last?

Saccades last between 20 ms and 100 ms - the length of the saccade is related to the distance that our vision shifts during the saccade.

Visual fixation durations in continuous visual tasks such as scene perception vary from less than 100 ms to several seconds.

The average look right / left lasts just under half a second. Half a second is 500 ms. The implication is that the eyes make relatively few fixations and relatively big traverses whilst scanning around the visual field.

The implication from that analysis is that the idea that we can't "miss an object entirely" is wrong, because the foveal zone is so tiny just a few degrees across compared with a visual field of approaching 180 degrees if we're emerging from a junction.

If we don't look in precisely the right place using our direct search AND we fail to detect the motorcycle in peripheral vision - at any kind of distance the head-on view of a motorcycle is just a degree or two across - then it's likely we'll miss is. Even if a fixation DOES fall directly on the bike, the brain still has to process the visual information to make sense of it in terms of shape, colour and silhouette to make sense of it before we 'SEE' it and this visual information becomes available for our CONSCIOUS brain to process a 'motorbike and rider'.

Fundamentally, because it's there in the visual field doesn't mean that it's going to be seen.

[Hot_Air]

On this theme, have you seen this video? Why our senses are fine-tuned for utility, not for ‘reality’

Anil Seth on why our senses are fine-tuned for utility, not for ‘reality’

It’s easy to mistake our conscious experience for an ongoing, accurate account of reality. After all, the information we recover from our senses is, of course, the only window we’ll ever have into the outside world. And for most people most of the time, our perception certainly feels real. But the notion that our senses capture an objective external reality can be dispelled by considering something as fundamental as colour, which can be culturally influenced and, even within a single culture, leave the population split between seeing the same picture of a dress as black-and-blue or white-and-gold.

In this instalment from Aeon’s In Sight series, Anil Seth, professor of cognitive and computational neuroscience at the University of Sussex in the UK, puts our imperfect relationship with reality in perspective. In conversation with Nigel Warburton, consultant senior editor at Aeon+Psyche, Seth argues that it’s not just that our perceptions provide flawed accounts of the outside world, but that our brains aren’t in the business of recovering the outside world to begin with. So it’s more accurate to think of our conscious experience as a series of predictions that we’re incessantly and subconsciously fine-tuning – a world we build from the inside out, rather than the outside in.

[The Spin Doctor]

On this theme, have you seen this video? Why our senses are fine-tuned for utility, not for ‘reality’

Nope but I have pointed people at the TV series and book by David Eagleman, called 'The Brain'. Well worth a watch / read.

(I got a 'like' from Dr Eagleman on one of my FB posts where I discussed his work )

[Horse]

So it’s more accurate to think of our conscious experience as a series of predictions that we’re incessantly and subconsciously fine-tuning

I'm no expert on neuro stuff. But

When you watch something that's moving, there must a delay for the information to travel from eye to brain and then be processed.

If you then decide to reach out and catch that moving object, you are predicting based on out of date information.

On a wide scale, as we move around, we're constantly making decisions to avoid other moving objects, similarly based on out of date information and predictions from it.

Perhaps:)

Edit: just reading one of Spin's books, it reminded me that there's an additional delay in the system. The brain sends the necessary messages to muscles to move the 'catching' hand - but there's a delay in message to action happening. So brain is predicting on out of date information then predicting where the catch will occur when the hand must actually be somewhere else.

[The Spin Doctor]

It's why we have to learn to catch...

...and why we trip up when we wear glasses we're not used to.

[Hot_Air]

All of which makes me wonder about overtaking: we need a lot of accurate predictions to get it right!

[iansoady]

The trouble is we rarely hear from those who got a high speed overtake wrong......