Tuesday, 6 June 2017

Urban mobility - harmonising platforms and infrastructure

Y’know, watching government regulators trying to keep up with the world is my favorite sport.”
Neil Stephenson, Snow Crash, 1992.
Technology metals and new materials offer the promise of a 'Cambrian explosion' in forms of urban mobility. Too many examples to list, but see the options at the end of this , or this. If we can co-develop infrastructure and mobility platforms in a functional way, then we may achieve remarkable levels of Quality In Use [1]. I have been unable to find any signs of work towards this aim, so this post has been written in haste  as a call for someone to point me in the right direction. It must be happening, surely?
Decent bicycle infrastructure was achieved in Denmark and the Netherlands only with a struggle; still to happen in the UK by and large.  Innovative approaches to bicycle infrastructure seem the right place to start, e.g. by expanding this, this and this.
The UK history of regulating innovative platforms is pretty patchy e.g. this or this on Segways and hoverboards, and this or this on microcars. Ebikes and pedelecs already seem a bit of a regulatory mess e.g. see this, this, this or this. Note also that speed is an important determinant of Quality In Use, and current standards may not be right, as discussed by Copenhagenize here.
For Système Panhard vehicles and their 20th Century derivatives, the Silicon Valley obsession with technology may not be a cost-effective approach (there are folk who claim a moral imperative to use AI to reduce accident rates - such folk are dangerous). Simple speed limiters might be better (though less popular).
Much of the current regulation seems arbitrary and appear to be based on (unstated) assumptions that are (or will become) very questionable, and use their own specialist language (invalid carriages, pedelecs etc.). They don't seem exoskeleton-ready. Modern platforms offer the potential to meet multiple regulatory categories at the press of a button or automatically. New types of platform need appropriate places in the infrastructure. At the time of writing, San Fransisco is considering a ban on delivery robots on the pavement (sidewalk there). Functional regulation is required to spare us from inappropriate regulations such as the urban myth of London taxis needing a bale of hay in the boot for the horse. This project between MIT,  the National University of Singapore, and the Singapore-MIT Alliance for Research and Technology (SMART) is worth a look. They converted a mobility scooter to operate autonomously. In two months.


Walkable urbanism


  • Accept that an integrated approach to platforms and infrastructure is the best route to safety, low cost access, and innovation. There is currently some very limited acceptance of multi-mode platforms for both invalid carriages and pedelecs/e-bikes, but way short of what is desirable. The potential for innovation may be best implemented with a major extension of multi-mode platforms operating according to the lane they are in.
  • Human Centred Design [2], prototyping inc. VR, AR, and consultation (Holmston Rd, Ayr I'm looking at you). Accept that designing for difficult use cases (disability, elderly etc.) benefits everyone else, and benefits difficult use cases by reducing costs. This code of practice may help (I have not examined it yet).
  • Regulatory capture [3] (e.g. Uber in London) to be treated with extreme prejudice. The public highway is to remain a public commons, and not to be privatised. Data and algorithms relating to the safe use of roads and pavements ditto.
  • Functional allocation of streets, and speed related lanes should lead to new opportunities for platforms with light regulation of design and operation for low speed platforms.
  • Accept that the US is an outlier in terms of urban design and public transport provision, and 'solutions' from the US should be treated with considerable caution, including their fascination with putting electric propulsion and advanced computing in 20th Century cars.
The approach to design implementation would seem to start with functional street categories, giving combinations of lanes. The potential use of new platforms needs to be aligned with an evolution of types of lane. A suggested arrangement is as follows:

Lane 1 - Pavement updated (UK pavement = US sidewalk); design speed 4 mph

Pedestrians, unpowered prams, buggies, trollies, carts, wheelchairs etc. Platforms up to two feet wide (legged, wheeled, hover - whatever) with a mode that limits speed (no licence, lights, horn etc required, but enough visibility and audible warning of approach, minimal regulation and liability), inc. autonomous PLatforms with or without people. Platforms without people need to behave appropriately e.g. for blind pedestrians. Platforms that can also operate in other lanes are fine here when in the Lane 1 mode.

Lane 2 - Cycle lane updated; design speed 10 mph

This is an average urban cycling speed and doesn't surprise other people. Having lanes 1 and 2 next to each other with just visible distinction seems to offer the most flexibility. Platforms with and without human assistance, with and without people. Platforms with no people will need some sort of official safety approval.

Lane 3 - Urban street updated; design speed 20 mph

If the functional design of the street has this as the upper speed limit, then lanes 1, 2, and 3  can be combined with no separation (cf. this), but platforms that travel at speeds higher than Lane 2 speed will need proper Type Approval, licencing etc.Lane 3 only platforms can be two people wide.

Lane 4 and above

Platforms capable of appropriate minimum speed, with suitable visibility, audible warning, protection.

Question - Folk must be working on harmonising platforms and infrastructure. Who is?


[1] Quality In Use is defined as:The degree to which a product or system can be used by specific users to meet their needs to achieve specific goals with effectiveness, efficiency, freedom from risk and satisfaction in specific contexts of use. ISO 25010 (2011)
[2] Principles of Human-Centred Design:
  • A clear and explicit understanding of users, tasks and environments
  • The involvement of users throughout design and development
  • Iteration
  • Designing for the user experience
  • User centred evaluation
  • Multi- disciplinary skills and perspectives
[3]  "When you try to regulate markets the first thing to get bought and sold are the regulators"  P.J. O'Rourke

Friday, 2 June 2017

Some reflections on 'Sully'

The movie 'Sully' received some criticism for taking artistic licence with the NTSB inquiry, and the NTSB complained about how they were portrayed. Additionally, some of the cockpit actions were criticised as incorrect - not according to procedure. This note rebuts such complaints and criticisms.
When I hear 'failed to follow procedures' this is the picture that comes to mind.

It is as important to examine appropriateness of the procedures as it is to examine the appropriateness of the crew behaviour.
There were no procedures for the situation they faced (See Airbus report here), so criticism of flap selection seems more than niggardly hindsight. This article has the following:
"The NTSB recommended changing the location of the rafts to ensure capacity for all passengers, since it's unlikely the rear rafts would be available. The FAA rejected that, saying that if Sullenberger had followed Airbus' directions on descent speeds for ditching, the rear rafts would have been usable. The NTSB said the ability of pilots to achieve those descent speeds has never been tested and can't be relied on. "
There are also questions as to the extent the investigation recommendations have been acted on.
As regards the NTSB moaning about being seen as adversarial, this from the scriptwriter has the ring of truth to it.
The key was, I had to do three layers of research," he says. "One was everything about the NTSB investigation, two was Sully's book...but then really the third level was memorizing Sully and Sully's willingness to share the stuff that he had not shared before - what he went through that was behind the scenes, that's was the wrenching and crushing investigation, the attempt, not out of ill will, but the honest attempt to try and find something that would affix blame. That’s really what they were looking for. You know, you look at 99 percent of these cases, the investigation, it always says at end, ‘pilot error.’ That’s the expectation even if someone is not going to speak that that's somewhere in the bloodstream of the investigation - pilot error. There was no pilot error to find. But it didn’t keep them from looking.”
Recall the press release for the incident report on Flight 447; it put 'human error' on the front pages of newspapers round the world (or at best 'pilot and technical error'). If you read this compelling analysis of the incident, a different picture emerges:
  • Two co-pilots flying rather than pilot/co-pilot, with #3 pilot as Flying Pilot.
  • The Air Data System froze (a known problem). Type Approval for Air Data Systems had not changed since the days of propellor aircraft flying at half the height and half the speed. This caused the Flight Computer to go into some sort of emergency mode.
  • None of this had been in the training and simulation for the pilots.
  • The Flying Pilot held the joystick right back; the other pilot would not have been aware of this, since the joysticks weren't coupled.
  • The hindsight interpretation of the stall warning appears to be controversial. It would appear that the manufacturer was keen to state that the situation facing the pilots was straightforward (i.e. human error) "The situation was not ambiguous and the stall was obvious,". The BEA investigators did not think matters were so straightforward, see here and here. Not surprisingly, there were 66 pages of discussion at PPRuNe
A remarkable incident of pilots vs. automation, where the pilots survived to tell their side of the story, can be found here.
To quote Sidney Deker 'Human error is a symptom of trouble deeper inside a system' - a consequence not a cause of accidents.