Connected and semi-autonomous vehicles currently run on over 100 million lines of code, but they’ll soon run on 200-300 million lines of code. This attests to the remarkable progress of automotive technology over the past few years, but it also underscores that there’s still a long road to the fully autonomous vehicle.
Vehicles with Level 5 autonomy will require an estimated one billion lines of code, far more than the semi-autonomous vehicles. After US$16bn in investments and years of predictions that fully autonomous vehicles (AVs) would hit the roads by the early 2020s, why does it remain a distant prospect?
Here are four issues that automotive manufacturers are considering to make self-driving vehicles a reality.
1. Technological limits
Self-driving cars will require technology – from processing power and connectivity to new, flexible and consolidated vehicle architectures.
Secure, high-performance data storage, processing and connectivity will be vital to supporting autonomous commands, over-the-air-updates, and vehicle add-ons like lidar sensors. The next generation of automotive zonal architecture will sufficiently support the features that are needed to power fully autonomous vehicles.
2. Holistic approach to infrastructure planning
Even if auto makers managed to upgrade vehicle architectures overnight, actually deploying AVs would require proper infrastructure – including vehicle-to-infrastructure systems like road and in-car sensors – which is not yet in place.
A contributing factor to this is that policymakers haven’t prioritized building the ecosystem needed to support AVs. While the US Federal Highway Administration recently announced US$43.3m in grants to 10 states for “forward-looking technology projects,” including for a freight truck automation corridor in Ohio and Indiana, measures to benefit passenger vehicles have been few and far between.
A uniform federal mandate to support AV infrastructure would move the needle, but the industry currently confronts a patchwork of conflicting local and state policies.
3. Securing the autonomous future
Once the industry manages to address the necessary technological aspects, it will be critical for policymakers and industry to build public trust in the safety of AVs.
By eliminating human error – the chief culprit behind road accidents – AVs could save about 10,000 lives per year according to conservative estimates. But while AVs could ultimately be beneficial to safety, high-profile accidents involving semi-autonomous vehicles have understandably fueled doubts about the safety of passengers and pedestrians.
Another major issue lies in how local governments would actually police fully autonomous technology and whom they would hold accountable for accidents caused by such vehicles.
Much remains to be done to address these concerns: uniform national safety standards will be essential, as will robust cybersecurity measures, given the sheer amount of connectivity in a car and hacker threats to AVs.
4. Human dilemmas
Philosophy’s famous trolley problem has a new spin on it. If an AV is inevitably going to crash and must crash into something, how should it ‘decide’ what – or whom – to crash into?
Naturally, people are uneasy about AVs making such a grave moral decision – and given that it will be human developers who ultimately program a car how to respond, this is an issue that merits serious discussion. That conversation has yet to begin in earnest – but deploying AVs without having had it would be irresponsible.
Even in less dire scenarios, AVs must grapple with tough human dilemmas. What parents would be comfortable allowing a robotaxi to transport their children to school? Leaving children in the hands of metal robots devoid of ‘human touch’ is disquieting to those speculating about AVs.
AVs on the horizon?
Contrary to the predictions of a few short years ago, Level 5 AVs won’t be populating the roads anytime soon – indeed, it may well be 2030 or later before they’re deployed.
Aside from the formidable obstacles already confronting AV development, the Covid-19 pandemic and the economic downturn have prompted auto makers to shift their focus to other types of vehicle technology with greater near-term prospects – including the promising electric vehicle market. Automotive technology may be advancing by leaps and bounds – but the fruits of that innovation won’t be seen in self-driving cars anytime soon.