Black box data could prove lynchpin to autonomous vehicle success

There has been considerable hype over the last few years about fully autonomous vehicles merging onto the roads, but the reality is that this still remains a way off from being realised. Electric and autonomous vehicles were once thought of as a dual phenomenon but their paths have since diverged: electric car sales are on the rise while autonomous technology has slowed, in large part due to the pandemic.

At this stage, cars have partial automation, but full autonomy—the kind described for example by Volvo in its short film showing a woman comfortably reading and snoozing while her car speeds along—is still in the distant future. But why is this?

There remains a lack of understanding as to what effect this automation is having on road safety. Before using fully or partially autonomous vehicles there is an element of trust required in this technology, which is new for most users. They need to be convinced that driving in vehicles with autonomous technology is safe. Risk compensation might be playing a factor in some incidents too, meaning people take greater risks when they feel safer or when they perceive that safety isn’t their responsibility alone. An element of human awareness and attentiveness is still required to circumvent a collision. Potential incidents are increasing the pressure on industry executives and regulators to ensure that automated driving technology can be deployed safely.

Black box audit trail

What if autonomous vehicles carried a black box, like an aircraft, that recorded a detailed account of everything that happened in the car’s systems as well as the drivers’ actions? This is not a theoretical question. As vehicles become less reliant on drivers and more dependent on automated technology, the interest among insurers, car manufacturers, governments, and regulatory bodies to capture data to facilitate investigations into incidents will grow.

Allowing various stakeholders with different agendas to look at the data would be beneficial in many ways. Insurers would want to gather data to understand the possible dangers associated with an autonomous vehicle before they provide coverage. Courts would like a data trail to help identify if someone associated with a crash was liable or not. Car manufacturers could incorporate learnings into their own autonomous vehicles, helping to avoid unnecessary dangers and propel the industry forward.

The ownership and accessibility of much of the data is in flux, as legislators and regulators play catch-up with the fact that human beings and AI increasingly share the driving. In the event of an accident, it would have to be clear in which way the vehicle’s autopilot or the driver himself or herself is to blame. Car manufacturers would have to share the vehicle’s data to clarify the question of responsibility. This raises ethical and legal questions about the valuable data stored in the black box. In the not-so-distant future it might become mandatory for car makers to share black box data.

What if autonomous vehicles carried a black box, like an aircraft, that recorded a detailed account of everything that happened in the car’s systems as well as the drivers’ actions?

Black box technology is going to add another element to the storage architecture, in that the sheer amount of data it records will need to be stored within the vehicle. The hardware used to store this data will require specific characteristics, such as shock absorption, as well as the intelligence to understand which data is crucial to save, since all of the data recoded can’t possibly be stored. Unlike the volume of data transmitted and received by in-car infotainment and advanced driver assistance system (ADAS) in autonomous cars—which is staggering and will only grow—data recorded in the black box can’t be stored forever. For example, the last 15 seconds of a trip may be all that is required to capture a collision. But the storage devices must have alignment with AI and machine learning to be able to manage that process in real-time.

Each plane in the sky carries a black box as standard, preserving everything and those recorders have been invaluable in working out the reason for a crash and advancing the industry by studying the data. The automotive industry could do well to apply these learnings to cars as autonomous vehicles move closer to a reality on our roads. Having the right storage device within the vehicle may benefit the many, not just the few.

---

The opinions expressed here are those of the author and do not necessarily reflect the positions of Automotive World Ltd.

Christoph Mutz is Senior Automotive Manager at Western Digital

The Automotive World Comment column is open to automotive industry decision makers and influencers. If you would like to contribute a Comment article, please contact [email protected]