As we move closer to self driving cars, a plethora of new challenges faces manufacturers and innovators. Currently, well over 90% of car accidents are caused by human error. However, as your car starts to take more responsibility than simply controlling the speed such as in cruise control, incidents caused by vehicle failure, error or even virus are likely to become more common.
While innovators are experiencing mixed results with the technology of the vehicles themselves, standard car manufacturers and their regulators are well ahead of the game. Blackbox technology has been mandatory in new cars since 2014 and can already offer a detailed analysis of any failure on the part of the vehicle. The advent of this technology, is of course related to a small percentage of computer failures with new vehicles. Referred to as event data recorders, these devices record just over five seconds of information before the crash, such as airbags, seatbelts and braking, and around one second after the incident – depending on the amount of damage to the vehicle.
Tuxera, a manufacturer of software for autonomous vehicles, aims to make of this technology far more advanced and bring it in line with aviation practices. CTO Szabocs Szakacsits, says the current methodology of analysing what happened immediately prior to an accident is fine currently but needs to develop rapidly. “There is a convergence, vehicles are becoming more autonomous and more and more sensors are integrated inside the car itself. This requires the black box to record, store and read the data, like distance to the next vehicle, traffic signs, road lights, and other objects surrounding the car.” Szakavsits mentions new issues such as power being cut just prior to a crash or during impact, “That’s why software components used in automotive applications need to become far more robust. It’s critical that the data is well-kept and failsafe, you can’t afford to have frame drops from the dashboard camera for example because that is critical information in case of accidents.”
As autonomous vehicles come online, the complexities won’t be limited to failures on the part of the vehicle you’re in. Accidents may result from other vehicles, infrastructure fallouts, and automation issues. As in the early days of aviation, it’s likely that incidents will occur at a higher level as the bugs are worked out. This will be further complicated by the transition period between normal and self driving cars – allowing for human error is a complicated task for the human brain, let alone the artificial intelligence of an autonomous vehicle. While it may seem grim, blackbox technology will speed up the process of innovation and ensure our self driving cars get safer, faster.