Cycling UK Forum

Pete Owens wrote:So to a certain extent that supports the position of minimising the transition period.

Thanks, yes. To me, the "minimze the transition period" argument is one that we might choose to deploy after AVs have been around long enough that we are confident it's the right thing to do. So the period might well be minimized by bringing forward its end date[1], but probably not by delaying its start date.

What I can't understand is worrying about the transition period itself. Every vehicle that is removed from the control of a slightly evolved chimp will represent an incremental improvement to our safety.

I can conceive of an argument that there would be confusion in a cyclist's mind about how a car is going to behave depending on whether it's an AV or not. But it seems weak to me, and I do think the "gradually safer" scenario is more likely.



[1] - And that might even turn out to be unnecessary. If AVs were a roaring success, but in 2050 still 1% of miles were being driven by humans, there would be a great deal of focus on those human miles. There would perhaps be zero tolerance for driving while distracted / tired / not an expert. It might turn out that human driving had thus also become a roaring success. Human driving might not end altogether until 2100, resulting perhaps in an 80 year "transition period"! (Obviously these numbers are purely illustrative.)

N.B. So I actually think the actual term "transition period" gives the wrong impression; as per my post at the start of this discussion.

So anyway...

Does anyone agree with Cycling UK's "CWIS" submission, where it states that the time during which there is a mix of human-driven vehicles and autonomous vehicles "will provide only disadvantages and no advantages for cycle use and cycle safety"? Can you elaborate?

Does anyone agree with me that my "gradually safer" scenario is also a realistic possibility?

Bmblbzzz wrote:Reading road signs might not be the ideal method but it's what various cars (presumably other vehicles as well) already do, or try to.

Confirmed. You can see a Tesla doing that / trying to do that in the video posted here: viewtopic.php?t=110566#p1071154

I took a screenshot at 0:37, and edited it with red graffiti to show a road sign apparently being detected http://i.imgur.com/EW1tkoL.jpg

landsurfer wrote:Fair enough .....

Thanks Landsurfer

landsurfer wrote:Item 2 and Item 4 are the same issues

No, they aren't.

Just to set things right .... i firmly believe in AV .... but i also firmly believe in getting it right ....

By derailing a conversation about item #4 with a host of other points? Even though Cycling UK has already been actively talking to the government about item #4 for at least a year?

I think this is getting a bit off track. (No pun intended.)

Landsurfer, perhaps the reason road vehicles are being automated before farm vehicles is because there's much more money in it? Take it up with Google, Nissan, Tesla, et al. if you really want to know the reason. And surely you aren't saying that Cycle UK should refuse to campaign about the specific aspects of autonomous driving that might affect cyclists, just because of your personal list of reasons as to why autonomous driving should/will be delayed indefinitely? I actually agree with you that IT security issues aren't taken seriously enough, but I think the powers that be will be more likely to take notice of a (non-IT) organization making that point if the organization was also getting involved in all the other issues too (especially the ones they considers us to actually be knowledgeable in).

So, lets imagine that a new campaign was added to http://www.cyclinguk.org/current-campaigns along the lines of "Vehicle Autonomy - Making Sure It's Done Right"

-----------------------------------------------------

Perhaps that new campaign might have a "Top 10 Issues" list, starting something like this:

1 - Autonomous vehicle behaviour in various specific scenarios, including passing distance, advanced stop lines, certainty of seeing a cyclist in all conditions, doors opening into traffic, etc..

2 - More focus on IT security issues by vehicle manufacturers and software makers.

3 - Due consideration of any problems that might arise from having multiple different autonomous driving systems all active on the public road network at the same time. (I put that in after reading The utility cyclist's comment above.)

4 - Due consideration of the problems that might arise during the period (however long it may be) in which some cars are driven by computers, and some by humans.

5 - Suitable prioritization of the safety of all road users. (C.f. Mercedes deciding to sacrifice pedestrians and cyclists)

6 - Strongly oppose the eventual possibility for "on-road cycling to be banned altogether" (quote from by Roger Geffen here).

7 - TBC, etc.

-----------------------------------------------------

Landsurfer, I hope the inclusion of issue #2 would mean you could potentially approve of such a campaign by Cycling UK.

But to be honest, that's not really relevant to the topic at hand. I started this thread to discuss issue #4, because Cycling UK have already submitted an opinion to the UK government about it, and as I explained above, I don't agree with that opinion.

landsurfer wrote:It is not the development of av road vehicles that should come first but the hardened, hack free software.
While software can be remotely hacked the whole concept of av should be on hold.

So from a perspective of what Cycling UK should be doing, it sounds like you agree with my proposal #1, and would want our new campaign to be very firm about IT security issues?

And what about my proposal #2? Perhaps it doesn't matter much to you because you feel the IT security aspect is so much more important?

A year ago, Cycling UK submitted a response about the Government's draft Cycling and Walking Investment Strategy (CWIS). Here's the PDF, but I'll quote the whole driverless car section here:

Driverless cars

The recent Queen's Speech included the announcement of the Government's intention to introduce a Modern Transport Bill, which will (among other things) facilitate the introduction of driverless cars.

The adoption of driverless cars could be either a huge benefit or a huge disbenefit for the promotion of cycling and cycle safety, depending on how it is done:

• It could lead either to large increases or large decreases in car ownership, depending on whether it makes car ownership normal among people who currently cannot drive (including children), or whether it substantially boosts the popularity of car-sharing clubs (enabling people to order a driverless car to come to their front door as easily as they can currently order an Uber car).
• Similarly it could hugely improve or worsen the safety and convenience of cycling. Fears that pedestrians and cyclists can simply run or swerve out in front of driverless cars – knowing that they would stop – could result in regulation that would severely restrict cycle movement. Conversely, the total adoption of driverless cars could mean motor vehicles being steered as precisely and predictably as if they were on rails. This would be extremely space-efficient, freeing up space for cycling. Further space could be freed up by reduction in car parking demand that would result from a marked shift from car ownership to car sharing clubs.

Cycling UK seeks assurances that driverless cars will not be permitted on roads other than motorways until it has been shown that they can reliably detect cycles and cyclists, and respond accordingly. Since cyclists negotiate for roadspace using eye contact, this is hugely important.

However, once the detection technology is reliable, we seek the rapid and complete introduction of fully driverless cars. We wish to minimise the time-period in which there is a mix of human-driven and driverless cars on the road, given that this transition phase will provide only disadvantages and no advantages for cycle use and cycle safety.

(my emphasis)

I commend Cycling UK for bringing up the issue generally. But I'd like to start a conversation about that final paragraph.

"This transition phase will provide only disadvantages and no advantages for cycle use and cycle safety." I accept that this is a possibility, but to imply that it is the only possibility seems wrong to me. I think there is another realistic possibility: that as soon as there are some autonomous vehicles on the roads, cyclists are safer, and as the proportion of autonomous vehicles increases, cyclist safety keeps on increasing. I refer to this as the "gradually safer" scenario.

Now as I said, this was a year ago. But I have also found a hint in 2017 that we may still have this mindset about a dangerous "transition period". Cycle-campaign-news-january-2017:

We also share concerns about the transition period.

If it does turn out that my "gradually safer" scenario is correct, then even the actual term "transition period" may be best avoided, as it might lead people to think in terms of everything already being planned in advance, including an 'endgame' for human-driven vehicles, and thus delay the potential early advantages of autonomous vehicles. (AKA self-driving vehicles)

My Proposals

1 - http://www.cyclinguk.org/current-campaigns A new campaign should be added, along the lines of "Vehicle Autonomy - Making Sure It's Done Right"

2 - The people within Cycling UK who are worried about a disadvantageous "transition period", should detail their concerns as part of this new campaign, but should not rule out all other scenarios unless they have documented a convincing argument.

Thoughts?