Autonomous vehicles — The hurdles awaiting ahead
Autonomous vehicles are one of the most exciting and revolutionary technologies our generation will see. Their path to our streets, however, remains full of challenges the industry has yet to solve.
Despite the illusion of immediateness created by the media and some stakeholders during the last few years, AVs have to surpass many hurdles for which the industry and society are still not ready and will still require some time to overcome. This post intends to provide and overview and some understanding on the nature and complexity of these challenges, while addressing how the industry is tackling them.
Technology
Undoubtedly, fully automotive technology is complex and will still require years of testing. However, promising advancements in software and hardware have proven sufficient for self-driving cars to navigate some of our cities with extraordinary results. In fact, these cars are able to drive around for millions of miles without human intervention of any kind.
Advancements in LIDAR, optics, object libraries, AI… are all proving that, technologically speaking, we’re not that far from where we’re aiming for.
Regulation — Ethics & liabilities
Efforts are starting to be made with tests in some major cities in the US, but do not forget that these vehicles count with a security driver, with whom ethic and liability dilemmas are “easily” solved. How our regulators solve for their needlessness is a tough one, which will require deep efforts from industry associations (e.g. AUVSI), universities and public bodies.
Players and stakeholders in the industry will have to work with regulators to develop applicable regulation that clarifies the framework within which fully AVs will be able to be deployed in our cities and, most importantly, co-exist with human-operated vehicles and pedestrians.
System divergence and the “tragedy of the commons”
Under the regulator’s and public opinion’s constant scrutiny, a fatality caused by the negligence of a single player could severely impact years of efforts of the whole industry. However, the high secrecy inherent to the IP of their technology makes it very difficult to implement best practices across the industry.
How can players protect their IP while at the same time help improve security industry-wide? Solving this arduous dilemma is key for the sustainability of an industry in which different paths can lead to very different (and fatal) results. Is it safer to use LIDAR or optic sensors? Should object libraries be open to everyone? Is it safe to gather information of ADAS or semi-autonomous driving from drivers? A reckless move of a single player could be fatal.
Funding
Time horizon for cash flow generation is far from short. During the years of boom of Silicon Valley with unprecedented liquidity in the market, funding has been abundant for players. However, post-COVID times seem less appealing for investors and there is a big concern about which players will be able to continue with their capital intensive operations in an investment shortage environment.
Even the projects funded by large OEMs, severely hit by the crisis, are likely to see their funding cropped. Possibly, consolidation and acquisitions will become a new trend, solving for some of the inefficiencies, but players backed by big pocket parent-cos now have an opportunity to claim and reinforce their leadership in the market.
Scaling up production
Will having strong manufacturing capabilities actually be a strategic advantage when the time comes to scale-up production? OEM-backed players will argue that it will evidently be, however independent or tech-company players will argue that AVs will never be built by the millions, especially in the dawn of the industry where deployments will be very gradual in the areas where regulation is favorable.
Customer acceptance
It is one thing to see a fully autonomous vehicle on YouTube and a very different thing to send your children to school in one.
Mobility service providers and AV companies have deployed programs in which users have used autonomous ride hailing services (e.g. Waymo + Lyft in Phoenix), but the security driver was still there providing a sense of safety. What will happen we he’s not there… that’s another story.
Cybersecurity
Can AVs be hacked? Short answer is, as any connected system, yes… but with a huge BUT. AVs are being designed by most players to be connected in the infotainment side, but processing systems, which rule the self-driving technology, are generally designed as an independent and autonomous node. For this reason, the vehicle does not require connectivity to do its vital function: driving. Therefore, hacking of the self-driving system should not be a threat.
Is autonomy the way to the promised land of profitability?
Despite the fact that self-driving technology is expected to solve the (un)profitability issue of ride hailing operators, there is a major shadow that casts over this prophecy: traditionally, price decreases in the transportation and airline sector have been passed to the customer due to increased competition, leaving near to zero value in the industry. Will autonomous players be able to lock down some of that value or will it basically erode their margins in the same way, providing extra value to customers?
Is it a winner takes it all market?
Despite competition in the ride hailing market has proven that local economies of scale are what determine the success in a specific region, AVs might change this paradigm. Should a player reach a point in which its technology is by far superior to competitors, compliant with regulations and should it be able to reach enough scale as to deploy in major markets, it might set difficult to reach standard for contenders. The barriers to entry might wipe most of the promising projects now underway, limiting competition and leading to an oligopoly.