January 11, 2022
Regulatory clarity, a push toward sustainability and explainable AI breakthroughs could be in store, the experts say.
Good afternoon! As we alluded to last Thursday, we'll now be in your inbox on both Tuesdays and Thursdays. In today's edition, we asked the experts to think about the future of autonomous vehicles. The path to their development and adoption hasn't been as linear as some of the early sweeping declarations about the technology may have suggested, so we asked the group what the milestone moments for AVs might be in the coming year. Questions or comments? Send us a note at firstname.lastname@example.org
Co-founder and CEO at Aurora
Deploying self-driving vehicles at scale means that the autonomous driving technology needs to handle edge cases that are challenging but rare — like a child running into a roadway, or an uncommon combination of vehicles merging onto a freeway. Since these situations can’t efficiently be happened upon by just driving more on-road miles, as autonomous vehicles edge closer to deployment, the value and importance of simulation will become even more obviously essential.
By investing early in simulation, companies like Aurora are able to leverage this powerful capability to safely create interactive scenarios for the system to both learn from and be tested against. Developers can simulate at a massive scale and quickly iterate on technology before it’s on the road. In fact, we estimate that our daily simulation tests are equivalent to driving approximately 47 round trips to the moon. Having access to this capability means that we avoid over-indexing on only the situations we see regularly, and instead can build a system that is broadly safe and robust.
In 2022, the investment will deepen and the value will become clear. Companies that don’t have simulation capabilities will need to invest in the difficult work of building high-fidelity, sensor-realistic simulation, without which addressing challenging edge cases will become prohibitively difficult. For companies, like Aurora, that have built the foundation, simulation will allow them to accelerate the delivery of next-generation transportation that’s safe, efficient, and autonomous.
Head of Policy at Nuro
2022 will be the year the autonomous vehicles industry truly gets serious about clean energy and sustainability. Last year, the state of California passed a monumental law — with support from Nuro, environmental advocates and other AV companies — requiring all autonomous vehicles to be electric by 2030, five years ahead of the state's target for traditional vehicles. We will see that future become even more of a reality in 2022.
With transportation continuing to be the leading cause of CO2 emissions, we have an opportunity and an obligation to transition away from using personally owned combustion engine vehicles for running errands to fully electric autonomous vehicles. At Nuro, we are designing and building fleets of 100% battery-electric AVs, and we are continuing to prioritize renewable energy and sustainability, including releasing our first ESG report later this week.
Last year, we also joined with other companies to found the SAVE Coalition, which is dedicated to ensuring AVs are EVs. From production to operation of these vehicles, the biggest milestones of the year will come from companies that demonstrate their full commitment to sustainability through the use of renewable energy.
Head of Commercialization for Trucking at Waymo
Since the start of the pandemic in 2019, global supply chains have been faced with a myriad of challenges, including a significant shortage of truck drivers. But this isn’t news as much as facts we’ve been living with. I believe that in 2022 we will continue seeing how autonomous driving technology can help alleviate these problems for years to come following promising results from last year. For example, in 2021, Waymo entered a collaboration with J.B. Hunt and expanded our UPS delivery partnership to include Class-8 trucks — hauling freight on behalf of both partners in Texas — where the Waymo driver delivered shipments safely and on time. In the coming months, we look forward to continuing to deliver goods in collaboration with leaders in the industry.
Investor at Alsop Louie Partners and Adjunct Professor at Stanford University
Consumer and regulatory trust remains paramount for the success of the AV industry — particularly for the machine learning and AI systems that power autonomous vehicles. To promote trust between man and machine, in 2022, technical advances should be expected in areas like explainable AI, robustness against adversarial attacks and reducing bias.
AV manufacturers have a responsibility to explain why vehicles get involved in accidents or make other traffic decisions, but the massive neural networks that underpin AI driving systems are difficult to interpret; R&D efforts will shift from simply making more accurate predictions to designing network architectures that are more easily interpretable by human users. Adversarial attacks can “fool” neural networks — for example, a bumper sticker with an optical illusion could be interpreted as a real 3D obstacle — and as such, automakers will need to partner with AI and cybersecurity experts in order to ensure their systems fare well even in the face of these occurrences.
Lastly, AV systems need to boost efforts to eliminate all types of bias in their models; for example, training data and test cases should address weather and terrain of various types, along with different types of scenes and obstacles that may occur in locales with different socioeconomic characteristics. Overall, as autonomous vehicles become fairer, smarter and more understandable, end user confidence will drastically increase, which should eventually translate into major real-world adoption on the road.
CEO and Co-founder at Embark Trucks
At Embark, we continue to believe the first large-scale commercial deployment of autonomous vehicles in the U.S. will happen in trucking. The recent technical progress in the autonomous trucking industry is a testament to that.
In 2021, we saw trucking companies begin to adopt and pilot autonomous trucks within their distribution networks. Major shippers like HP, Inc, AB InBev and top national trucking companies like Knight-Swift, DHL and Werner Enterprises partnered with AV companies including Embark in preparation for commercial deployment of AV trucks in the near future.
This year, there won’t be one singular achievement that defines the AV industry. It’s about building layers of innovation over time versus focusing on one specific moonshot. In 2022, we’ll see the gradual path to commercialization continue through a variety of technical and business achievements.
We’ve previously outlined some of the challenges that represent hurdles to commercial deployment of AV trucks — they include things like interacting with emergency vehicles, dealing with blown tires and even navigating snow. I expect that Embark and other players in the space will make progress toward building capabilities to surpass those hurdles.
2022 will also mark a crescendo of interest and investment from the trucking industry, as fleets will eventually be the end users of autonomous truck technology. In 2022, the group of early adopters will grow, and fleets will prepare for autonomous trucks by completing detailed network analyses to optimize their networks, get hands-on with autonomous technology and trial autonomous trucks in their real-world operations.
Deputy General Counsel and Vice President for Policy and Regulation at Luminar and Former Chief Counsel at the National Highway Traffic Safety Administration (NHTSA)
2022 will see the deployment of next-level autonomous features as well as the entrenchment of certain trends. The development of better and less expensive autonomous hardware and software has seen exponential growth fueled by greater field experience and smart investment by equipment suppliers and OEMs. In 2022, consumers will have better choices of autonomous features, and notably higher-levels of autonomy, that are supported by a more demonstrable and stronger safety case with key-enabling technologies coming to market. For example, Volvo's next-generation XC90, to be revealed in 2022, will provide this higher level of autonomy, including an unsupervised autonomous driving capability for highways.
I believe the next-generation XC90 will also lead an important trend among automakers that focuses not as much on the subscribers of cutting-edge tech but rather the real value case for customers in terms of safety enhancements, time returned to the driver, and operating stress eliminated This value case is key to mass acceptance and deployment — and brands have gotten the message. Finally, I think we'll also see a continued movement away from the discussion of robo-taxis, or a “new car ownership model,” and to a further concentration by suppliers and OEMs alike on enhancing safety features and enabling highway autonomy.
CEO at SafeAI
The AV industry in 2022 won’t be defined by a single milestone or piece of news, but by real-world traction. We can expect meaningful progress in three areas of autonomy.
First, on-road. Faced with complex driving conditions and regulatory challenges, self-driving cars have taken time to come into their own; this year, we’ll see real progress. Personally, I expect significant advancement in Tesla’s full self-driving features, approaching Level 4 autonomy by the end of the year and setting an ambitious tone for the next chapter of on-road deployment.
Off-road, I’m even more bullish. Labor shortages, dangerous working conditions and rampant inefficiencies have created huge demand for autonomy in off-road industries like agriculture, construction and mining. And with tightly-controlled environments, repetitive work and fewer regulatory hurdles, these industries are ripe for immediate disruption. Last year, we saw major companies like John Deere enter the fold with investment in off-road autonomy; this year, we’ll see a significant increase in connected, autonomous sites.
Finally, I believe we’ll actually see the greatest acceleration in perhaps the least-talked-about facet of autonomy: the supporting tools and infrastructure. The technology that enables autonomy — simulation tools, data infrastructure and processing, etc. — often flies under the radar in mainstream conversations, but these tools provide the force behind both on- and off-road players. Companies in this space are already successfully generating healthy revenue and valuations. In 2022, as applications for autonomy grow, these companies will continue to increase revenue and cement themselves as core players in the industry.
CTO at Velodyne Lidar
While the timing is uncertain for when autonomous vehicles will be a reality, 2022 will focus on advancing regulations around AVs, helping to create safer experiences on the road. From automakers to policymakers, we expect an evolution in the implementation of full-stack sensor suites, including hardware such as cameras, lidar and radar, as well as software, as officials work to provide a solution to growing traffic collisions and fatalities. Right now, some in the industry are trying to figure out how to get by without lidar, even though last year we saw that camera-only AV solutions do not often work in silo. Unlike cameras utilized by companies like Tesla that produce 2D images of the environment, lidar “sees” in 3D. Whereas cameras have to make assumptions about an object’s distance, lidar produces and provides exact measurements. With accuracy and safety in mind, a lidar-centric autonomous system is a necessity.
Additionally, as electric vehicles become mainstream, thanks in part to the new infrastructure bill, AV and EV adoption will grow in tandem. This means we'll also see the enhanced need for sensor technology that prioritizes power efficiency and performance while also maximizing range.
Velodyne is working closely with policymakers to consider lidar solutions in all AV regulation and standards moving forward, as lidar technology is a critical component for a safe, autonomous future.
See who's who in the Protocol Braintrust and browse every previous edition by category here (Updated Jan. 11, 2022).
Kevin McAllister ( @k__mcallister) is a Research Editor at Protocol, leading the development of Braintrust. Prior to joining the team, he was a rankings data reporter at The Wall Street Journal, where he oversaw structured data projects for the Journal's strategy team.
More from Braintrust