Self-Driving Teams Have Always Strived To Measure Safety. What If That’s Not The Hard Thing?

Topic: 
Tags: 

In the robocar world, everybody is safety-obsessed. But what if what's holding things up isn't that, but the fact that focus on safety had delayed the good road citizenship needed to operate a real service. Is good road citizenship even harder than safety? What ways might we measure it and get the trade-off right. I discuss this in a new Forbes site article seen in:

Self-Driving Teams Have Always Strived To Measure Safety. What If That’s Not The Hard Thing?

Comments

I posted the following article about a year ago on LinkedIn.
I think road citizenship is like the Turing Test. Imagine looking down at the traffic and not being able to tell which vehicle is being driven by a good human driver and which vehicle is fully autonomous - that's how good it needs to be.

THE FOUR HURDLES TO ROBOTAXI DEPLOYMENT
by Dr. Chris Borroni-Bird

Assuming that cybersecurity, ethics, liability and regulatory concerns can be solved for auton-omous vehicles (AVs) in general, the specific use case of geofenced L4 robotaxis will need to overcome these 4 specific hurdles before widespread deployment can be assured.

HURDLE 1: SAFETY
Safety is the first priority and people need to be convinced that AVs are actually at least as safe as typical vehicles. Acceptance will, inevitably, take time and familiarity. Although cyber-security concerns "lurk in the shadows” it is commonly accepted in the technical community that AVs should be safer because they have better sensors, are never distracted or drowsy and obey all road rules. However, for now, acceptable safety requires nice weather and cau-tious vehicle operation, which brings us to the second hurdle.

HURDLE 2: ROAD ETIQUETTE
A robotaxi can obviously be made safer if it moves slowly but this is a low bar to set. Ultra-cautious operation will be unacceptable to other road users and slow down traffic, which could lead to road rage and greater resistance to robotaxi operation. Moreover, passengers “stranded” inside a robotaxi while it hesitates to negotiate a challenging merge or intersection will think twice about hailing another robotaxi. Road etiquette still has to be proven but if we assume that robotaxis can be developed to operate as seamlessly in urban traffic as a regular taxi then the third hurdle facing their deployment is going to be local public policy.

HURDLE 3: LOCAL PUBLIC POLICY
It is increasingly recognized that the recent increase in ride-hailing has caused more traffic congestion in city centers. This should not be surprising because, historically, whenever mobil-ity has been made easier (cheaper, faster, more convenient) it has induced more demand, leading to more vehicle miles traveled and energy consumption. Building more roads is not a practical solution so cities may require robotaxi operators to pay a congestion tax if they drive around with nobody, or even just one passenger, in them, and may require them to serve less affluent populations and areas of the city and to improve accessibility to transit. This quid pro quo will negatively impact the business model, which brings us to the fourth hurdle.

HURDLE 4: BUSINESS MODEL
Ride-hailing companies today do not pay for vehicles, fuel, parking, maintenance and cleaning because the driver pays instead. Robotaxi operators will have to cover these costs and more. For example, robotaxis have expensive hardware (e.g. lidar sensors, lithium ion batteries) with uncertain durability and they use more energy (e.g. power draw from the computer, more vehicle mass and aerodynamic drag). They will also have to fund the charging infrastructure assuming they use battery electric vehicles. In short, much more technical and commercial work lays ahead before robotaxis will operate at scale in cities around the world.

Yes, all 4 of these have been worked on an discussed extensively. And like your list, people have always said "safety first." The question I open up is what if #2 is the harder problem, and you need them all.

There are people saying that #4 is the hard problem, actually, which I will write about later

Add new comment