**This post contains affiliate links and the publisher may be compensated if you make a purchase after clicking on these links.**
Today we look at the last of the four major automotive innovations reviewed in this series: autonomous vehicles, also known as self-driving cars. This was placed at the end of the series because it is also the innovation that will be the latest in coming, and especially in becoming dominant, if it does. As a reminder, the other three innovations we have covered are Vehicle Connectivity, Mobility Solutions, and Fleet Electrification.
Autonomous Vehicles Overview
Almost every vehicle on the road today has some autonomous features – the first one available was cruise control – but with features like parking assistance, lane departure warnings, more advanced cruise control, and collision avoidance features new vehicles are more autonomous than ever. Beyond that there are several companies, including Waymo, part of Alphabet (nee Google), that are developing and testing self-driving vehicles. As we get closer to some of the dates that were thrown out awhile back for when we would see fully autonomous vehicles it is clear that the vehicle driven by a human, and the need to continue testing and licensing drivers, is not all that close to ending.
There’s a term – V2V – meaning Vehicle to Vehicle that is at the heart of the real need to make self-driving cars a reality: the ability for vehicles to communicate with one another. The various self-driving cars are marvels of engineering, and the tests completed so far are testament to that, but there are limits to dealing with some elements that these cars would face in the real world: bad weather, bad roads, and worst of all (and most prevalent): bad drivers. Obviously I don’t mean you; I mean the people you call idiots and morons (and whatever else) during your daily adventure on the highways and byways.
Self-driving vehicles will continue to improve, however it is quite possible that you’ll see them first in specific places where they don’t need to compete for space on the road with traditional vehicles (like dense city centers: Manhattan, City of London, Shanghai, etc.) because traditional vehicles would be forbidden or rerouted. Additionally, placing them in these concentrated areas would create a specific transportation option people could access: effectively a taxi or car service, but without the driver. (Possibly it could be integrated in with Twitter and a text-to-voice app that will read tweet after tweet of half-baked ideas on what’s wrong with this country/city/local sports team, and what can instantly fix it, for folks that want the authentic taxicab experience.)
Similar to the other items mentioned, autonomous vehicles are as much a technology challenge as a legal and regulatory challenge. When we move to autonomous vehicles there will still be a need for auto insurance, as well as vehicle registration and titling to establish an owner and operator. If the passengers are not operators (i.e. drivers) at all it may be that leasing is the best option, especially at the outset of fully autonomous vehicles getting on the road, because it’s not financially smart to accept a liability on something you cannot control – which will be the position the owner would be in – so leasing directly from the manufacturer, distributor (i.e. dealership, but likely larger than today’s single point dealerships), or leasing company that will likely be affiliated with either the manufacturer or distributor will prevent this discounting in the consumer marketplace to account for the risk.
Solving for the Risks
System Malfunction Risk
The first of these risks is the possibility of system malfunction. It is actually easier to design an entirely automated vehicle where the user’s input is limited to entering a destination address and then letting the vehicle figure out the rest. In fact, the Google (now Waymo) self-driving vehicle did exactly that. There’s no need to design steering components, nor accelerator or brake that connects between the cabin and the related systems that move and stop the vehicle – because the computer will do that all without input from the passengers. (To be clear, Waymo presents on their website that there are not “pedals”, it does not say there is no user input at all). In any event, the design of the fully autonomous vehicle must have some plan for what will happen if there is a failure of key systems related to safely operating the vehicle, and it cannot be as simple as “if a key system fails, apply the brake”. That wouldn’t work very well when the vehicle had been moving at highway speed on a busy highway where the other vehicles are also moving at high speed and wouldn’t expect a vehicle to suddenly brake. Also, obviously if the system that fails is the braking system, there needs to be some other plan in place.
While this risk of system malfunction can never be eliminated, reaching a point where this risk is considered remote will likely be necessary given the high potential severity associated with this type of risk. Given how most vehicle accidents today are the result of human error, we have probably already reached a point where autonomous vehicles can perform better than human drivers. However, continuing to reduce the risk through increased reliability of the components and system redundancies that are designed in to prevent actual failure will almost certainly be the way to overcome this risk.
In the case of the vehicle specific systems, the process for accepting new software updates would need to be controlled so that malware couldn’t be loaded into the main systems: there would be problems if the systems for moving the vehicle had the software reprogrammed to accelerate when there’s a red light and to stop when there’s a green light. But the concept of determining whether to accept a software update isn’t completely new: your computer does it today using your anti-virus software, so this concept would simply need to be adapted to the way your vehicle would get new software updates. Besides the possibility of a malevolent third-party trying to hack into the car’s systems, this software security process will also need to prevent the vehicle owner from loading software that changes the main functionality. We unfortunately live in a world where vehicles have become a weapon when people intent on causing harm have driven into crowds. There will need to be substantial security in place to ensure an autonomous vehicle cannot be programmed to do the same by unknown assailants that can strike without recourse.
The network would also need to be protected so that incorrect signals are not transmitted to the vehicle’s sensors, which could in turn trigger improper actions to take place. In a world with only fully autonomous vehicles on the road we likely wouldn’t have the traditional traffic lights of red, yellow/amber, and green lights to control traffic through an intersection, but could likely rely on the vehicles receiving electronic signals directly to either stop or proceed through the intersection. If the network controller that replaced the traditional traffic light gets hacked, it could give the wrong signals to vehicles, so these devices would need to be secured as well as the vehicles themselves. Similar to this risk, the vehicles would need to receive some sort of “key” with each signal transmitted so they know it is from the real controller and not a fake one setup by hackers to disrupt safe vehicle operation.
What is Means for Leasing
Fully autonomous vehicles will not automatically change how vehicles are owned: presumably there would still be the option to own an autonomous vehicle (including using financing to buy it) or to lease it. Looking at other examples of new technologies, it may make more sense initially for all the autonomous vehicles to be leased. Almost all the risks of failure, and potential liabilities associated with it, should come back to the manufacturer of the vehicle, since any accident would potentially come back to their failure. For that reason “owning” an autonomous vehicle may be legally imprudent because it sets the owner up with liability they may not deserve to undertake (and insurance companies would charge accordingly, or potentially refuse to insure initially). Also, as mentioned, one of the risks to overcome is the vehicle being “hacked”, but this could include the owner placing malware in the vehicle’s systems because of planned misuse. Leasing will provide a basis for preventing the registered user of the vehicle from having any right to alter the vehicle software or systems. We will see how this actually plays out, and the impact of the other innovations discussed early may impact this as well.
Next Article: “Innovation Series Summary”
Previous Article: “Vehicle Electrification“