Autonomous cars: Saving me or saving you?
Transport technology

Autonomous cars: Saving me or saving you?

From road safety and software beta versions to ethics, morals, and logistics. Latest developments in autonomous driving and why we should not wait too long.

Autonomous driving is expected to become a reality more quickly than originally predicted. A swift implementation would certainly be considered a disruptive change. Self-driven car-sharing vehicles – ordered by app and arriving on site within 5 minutes – could revolutionize urban traffic. Driver shortages and driving time provisions in the area of freight forwarding would be a thing of the past.

The fatal accident

But now the first known fatality has hit the media. It occurred during a drive on autopilot in the US – with a car from Tesla Motors. What’s the impact of this accident on the continued research, consumer perception, and timely implementation of driverless systems?

Some numbers: so far, Tesla cars have covered over 200 million kilometers on autopilot before this fatal accident happened. On global average of all vehicles, there is one fatal accident for every 95 million kilometers covered. And another statistic cites that nine out of ten accidents are caused by human error – the greatest risk in road traffic is the human.

On autopilot, drivers are required to keep their hands on the steering wheel at all times. They are fully responsible for their own drives. Tesla’s steering system is still at a preliminary stage and as such, it is not mature enough yet for fully autonomous driving. Tesla CEO Elon Musk also conceded that the software still struggles with the recognition of pedestrians or cyclists.

Musk now has to take a lot of criticism for pushing the hype and driving aggressive marketing campaigns. It may have set the expectations bar too high and blurred the rather difficult path ahead from vision to reality.

Software beta versions

Should we promote and deploy beta version in this manner in real life? Can it be justified to suggest more than there is with targeted naming of things (autopilot) and putting human lives at risk through the pretense of wrong facts? Certainly not.

The expression “beta version” is supposed to make it rather clear that a preliminary version is meant and not a finalized software product. Customers act as test pilots and support the development of the final product – or its improvement – through experience and feedback. The continuously increasing application of agile software development methods promises higher flexibility, faster implementation, and related economic advantages.

One could form the opinion that the safety of people has been sacrificed on the altar of profit, and hidden under the cloak of innovation. The license for this Tesla car, by the way, was granted by the Dutch vehicle licensing authority. A speaker of the Federal Motor Transport Authority in Germany commented that the local authorities would not have granted a license to a system with “unfinished” software – as published on PC Games Hardware. Publicly naming and shaming the general use of beta versions, however, is a similarly suggestive approach that ultimately hurts the entire industry sector.

Nothing is set in stone

To make it clear: There is no such thing as finished software. At the most, there are versions that are commonly understood to run in a stable manner and to offer functional consistency. In a world of constantly changing processes and continuously advancing technologies, there are no conclusively final systems. Just as in the very same way, there are no processes set in stone. Not only Musk clearly recognized that such software-based systems require continuous updates.

In the software industry, we speak of “continuous delivery” as a collection of techniques, processes, and tools for the automation of implementation and delivery processes. Tesla might be a trendsetter here, but adaptations via cloud technologies will soon establish themselves as a standard in the industry as part of the increasing digitization of cars.

Safety-critical equipment in a car, however, requires much higher standards than software for entertainment electronics. And current performance levels of a system must be unambiguously clarified, too. The German automotive industry developed a five-stage model for autonomous driving this way – from cars fully guided by the driver to fully autonomous vehicles that require no human intervention. At the moment, the automotive industry is at the second stage – you can find more details in the EU Briefing from January 2016.

Is autonomous driving dangerous?

No, quite the contrary: it is rather likely that achieving the ambitious goals to reduce traffic fatalities will depend on autonomous driving. So the question is, how safe does the technology need to be in order to be better than a human driver? Human drivers are much more prone to driving errors, impairment through emotions, and traffic judgement errors than computers. We may expect a drastic decrease in accident numbers following an area-wide implementation of autonomous and integrated driving.

Based on various surveys, this view has not yet gained full public acceptance. DEKRA – an internationally active service-providing company – introduced some results at last year’s IAA Commercial Vehicles exhibition, featuring public opinions in the US, New Zealand, France, and Germany. Germans seem to be most skeptical, with only 8% believing that fully autonomous cars will catch on within the next ten years. On the safety issue in particular, however, at least the majority of respondents believes that automation will have positive effects indeed.

Source: Forsa survey for DEKRA
Source: Forsa survey for DEKRA

Legal framework. And ethics and morals.

The development of the new legal framework to include all specifics of fully automated vehicles will be much more difficult though than any technical challenges. This will require a comprehensive change of thinking: responsibility will be shifted from the individual to the public.

What will this mean for the software and its interpretation of the most classic dilemma-scenario of autonomous vehicles: Should the car be programmed against the driver and in favor of the safety of crowds? The majority of people are in favor of such moral behavior – as per a research study by the specialist magazine Science that was released in June. At the same time, however, the study’s results also reveal that the majority of respondents would take their personal purchasing decision in favor of a car that prioritizes their personal safety.

German Transport Minister Alexander Dobrindt plans to found an ethics committee, which will work on these – and other – ethical questions around driverless cars. You can read more about it on “The Strategist”, as published this month.

A decision on the described dilemma is required, but it will also delay the success of autonomous driving. And with this delay, another dilemma will arise: the later autonomous driving will advance, the longer we will have to face the high numbers of traffic fatalities. With this in mind, it’s rather sad to hear that we need more time to find better answers.

And not to forget about logistics!

Beyond the public and personal aspects of this topic, it’s also about potentials for logistics. Unmanned transport vehicles only form one of the building blocks in the comprehensive process chain with minimal human interaction. In addition to our opportunity to reduce the number of accidents, we may also expect a reduced environmental impact with gains in productivity at the same time. So, the chances clearly outweigh the risks.

What are your thoughts on this? I very much look forward to your comments and feedback on LinkedIn.

In the meantime, please explore our portfolio of software solutions to modernize logistics and global trade processes in your supply chain and sign-up to our newsletter to stay up-to-date on industry developments.