Looking Ahead: The Future of Regulations on Autonomous Vehicles (Industry Perspective)

Autonomous vehicle technology progressing rather rapidly. So much so, in fact, that until the U.S. Department of Transportation (DOT) announced its new AV policy, technology development was outpacing regulatory requirements.

Now, however, the technology and the regulations seem to be more or less parallel in terms of development. But it’s important to take a closer look at this technology and the regulatory environment in which it will evolve, because before we know it, Ford, Google, Tesla and Uber will be putting more autonomous cars on the road in more and more places. So to protect public safety while allowing this sector to develop, policymakers will need to look out beyond where the technology is today. 

Leading up to the recent DOT ruling, automakers said they wished to avoid “a patchwork of state laws” here in the U.S. This goal seems to have been sufficiently achieved, as the DOT policy essentially says that the state rules of the road will apply while a human is driving, but when the computer takes over, so does federal policy. The primary reason for this is safety: Just as we have one federal standard for lifesaving features like airbags and seatbelts, we’ll have a uniform federal standard for autonomous vehicle safety.

The DOT policy is, first and foremost, focused on safety — and with good reason. Indeed, one of the biggest promises of autonomous vehicle technology is its potential to dramatically reduce the number of vehicle-related accidents and fatalities. Safety should absolutely be the chief concern of federal regulatory bodies.

The policy also is focused on addressing questions like, “What happens if a crash happens?” and, “How do we piece together the different data points after the fact and try to figure out why the crash happened?” Or if there was a near-miss scenario, how will we utilize the data from that near-miss scenario to make sure that the collision threat doesn’t happen again? In addition to data recording — sharing this data for real-time driving situations — safety regulations are vitally important. 

DOT officials are thinking very hard about this, and they're thinking from a different point of view than the technology industry is. While the tech industry has been thinking about safety, their initial focus was on technical safety issues — issues like which sensor is the best to handle a particular use case, and will LIDAR, RADAR or an optical sensor be the best way to detect a particular object? Government has been looking at arguably less technical — but equally important — issues.

Of particular importance is the fact that the tech industry will be looking at and contributing data on a variety of uses and scenarios, as well as expressing opinions and sharing what is learned with  government — a necessary and significant collaboration.

Overall, this DOT policy is definitely a step in the right direction. In order for the autonomous driving sector to become a safe and widespread reality, there are a few more things to consider on a continual basis as this technology evolves.

1. Testing in different climates and traffic conditions. Thus far, real-world testing of autonomous vehicles has primarily taken place in areas of California and Nevada, where adverse weather is not a serious or regular concern. In order to obtain widespread adoption and a higher degree of safety, testing in bad weather and different climates (and more dynamic traffic situations) will have to be done, and appropriate changes to both the technology and applicable regulations will have to be made. 
2. Wireless upgrades. Given that the future will require autonomous vehicles to communicate with each other, a higher degree of speed and reliability in wireless communications will have to come to fruition. 5G wireless is expected to meet these requirements, but we’re still at least five to seven years away from widespread 5G deployments. This area will just take some time, and regulations will need to evolve along with the technology. 
3. Autonomous vehicle regulations. There are significant differences in traffic laws (and even direction of travel) from country to country around the world. Not only will individual countries have to iron out their own regulations for autonomous vehicles, but there will have to be appropriate degrees of cooperation among countries, especially those neighboring countries around the world whose traffic laws differ. The cars will have to be technologically capable of adhering to local laws regardless of their location, adapting as they travel across borders.
4. Data, data, data. Machine learning will improve over time, and we simply need a much larger volume of data than we have right now in order to improve the ability of these vehicles to make better decisions in real-time. Every mile driven by autonomous vehicles and every piece of data acquired by every sensor will add to this body of knowledge, thereby improving our ability to develop appropriate safety regulations, which will help lead to better accident avoidance and prevention. 
5. Tech standards, regulations. Since a vehicle’s ability to make split-second decisions is only as good as the data it has available to it at the time of a potential accident, we will need strong regulations in the area of sensor quality and proper functioning. Autonomous vehicles will need to be able to self-certify that all of their safety-related sensors are in proper working order at all times. 

Serhad Doken is the vice president of partner devleopment for InterDigital.