In the final analysis, the definitive question in assessing autonomous vehicle operation is, are you the driver or are you a passenger?
Clearly, there has been a significant increase in the availability of automated driving technology and the industry continues to develop and market new and enhanced systems. In order to provide a common vocabulary when discussing emerging technologies, SAE International (formerly known as the Society of Automotive Engineers) has introduced a five level system.
Level 1. Driver Assistance: This describes many of today’s new cars. The driver is responsible for the safety and operation at all times, but the car can take over at least one vital function, such as steering or speed control. Adaptive cruise control is the best example of existing technology.
Level 2. Partial Automation: Today’s more advanced cars have reached this level. The driver is still fully responsible for the safe operation of the vehicle, but it can take over steering, braking and speed under certain conditions.
Level 3. Conditional Automation: The car has the ability to control acceleration, braking and steering. It can stop at stop signs, obey traffic signals and signs, and navigate complex traffic situations without driver intervention. Although the car can drive itself, the driver must still pay attention and take over, instantly, at any time. This level still requires drivers to be fully ready to take over, and it ultimately holds the driver responsible if an accident occurs.
Level 4. High Automation: This level essentially means that the car is capable of full automation, but it can give drivers a warning and time to take over control. Level 4 is fully autonomous, allowing drivers to do something other than pay attention to the road.
Level 5. Full Automation: At this level there is no human driver and automakers envision vehicles without a steering wheel, accelerator or brake pedal. The only difference between level 4 and level 5 is that drivers will not have the ability to take control at any time.
All of this raises some interesting challenges. How do you anticipate consumer preferences, particularly given lead times and rapidly changing advances in technology? Also, and perhaps even more important, how do you assess liability in case of an accident? Experiments have shown that users of partial and conditional automation systems put too much trust in technology and stop paying attention. Will, “the car did it” prove to be a viable defense?
All comments and suggestions are welcome.
Walter J. Kirchberger, CFA