Mapping race to autonomous vehicles

• Autonomous Vehicle (AV) industry and the growth story
• Relevance of HD maps with respect to AVs; challenges and potential payoff
• Varied business models being practised in the industry domain
• Possible challenges posed for the industry

The autonomous vehicle revolution is set to disrupt all technologies known to man till date. In this rather exciting time for humans, the reel life sci-fi films have made their way to the real life. An autonomous vehicle (AV) or a driverless car is one that is being self-driven from a starting point to predestined destination and functions on an autopilot mode. In order to enable this, various technologies, as well as sensors, are placed including the likes of active steering, cruise control, GPS navigation system and anti-lock braking system. Besides offering a comfort of hassle-free driving, KPMG expects AVs to result in a reduction of accidents by approx. 80%, which are primarily caused by human negligence.

While partially automated technology has been in existence for the last couple of years, fully-driverless cars are currently in advanced testing phase. In addition to AVs, their electric versions are the perfect recipe to overhaul the automobile sector as a whole. BCG expects more than 12 million plus fully autonomous vehicles and 18 million plus partially autonomous vehicles to be sold each year by 2035. Further, up to the middle of 2017, a total investment of approximately $77 billion has been done in the industry, while, the cumulative mainstream investments accounted for $75 billion and the cumulative start-up investment accounted for the remainder $2 billion.

In contrast to the traditional automotive industry, the AV industry has undergone a massive transformation with the traditional ecosystem being challenged by evolving business models and technological developments. Moreover, several start-ups and established industry players have entered the industry to reap profits by being the building blocks of the industry. These include traditional car manufacturers (OEMs), mobile phone manufacturers, silicon chip designers, digital cartographers (mapping companies), graphics card manufacturers and internet giants among several others.

While numerous aspects contribute to the success of the AV industry, maps form the base for the development and its take off. In order to enable flawless crash-free driving, AVs require updated maps for each and every conceivable route possible. Further, for a robot-driven vehicle, one of the key components of the drive is not only what it sees, but more importantly, what it knows beforehand about the area it is travelling through. In short, a car must understand the road environment beyond the range of its onboard sensors. This requirement has essentially led to a full-blown mapping war. Furthermore, several companies have jumped on to this bandwagon and are spending billions to cash in on a staggering payout that could still be a few years away. The Goldman Sachs expects the mapping market for AVs to grow in value to a whopping $24.5 billion (2050) from $2.2 billion (2020) as seen in the below chart.

Among the majors, one of the prominent names that come up is that of Google, which aced the traditional digital maps years ago with a superior software expertise and its streetcars blanketing the globe. The company boasts of more than 1 billion users of its smartphone app every month after beating all other navigation companies, automakers and even the smartphone giant Apple Inc. Google’s prevailing dominance in digital maps gives it a head start in HD maps, which will be primarily accruable to Waymo, its AV spin-off. The company has established partnerships with Avis, Lyft and Fiat-Chrysler. Further, apart from Google’s very own real-time mapping data, Waymo will offer its customers sensors and software which, in essence, form the heart and brain of a robot car. The competition, this time around, is again fierce and companies are trying different business models to race ahead of each other. A few OEMs (Original Equipment Manufacturers), for instance, are choosing to collaborate with technology companies. In spite of being competitors, German automakers BMW, Audi and Daimler are sharing their sensor data via a cloud-based service provided by HERE. On the other hand, Volkswagen and GM have opted to lean more on third-party services in order to have their maps off the ground. For example, Mobileye, a self-driving car start-up has already inked agreements with BMW, Volkswagen and Nissan for a mapping product. The aim as per Dan Galves, Mobileye's Chief Communications Officer, ‘is to pool together mapping data from car manufacturers to generate as much scale as possible’. AV enthusiast, Tesla has taken on a different approach. The company has been using Autopilot: a driver-assistance software, which relies on sensors and cameras on its fleet. Tesla has leaned on Mapbox Inc. (backed by Softbank) to help it assemble its maps and paid $5 million for a two-year licensing deal.

Additionally, there are several small start-ups that are operating and are hawking specialised software or gadgets to develop these maps for OEMs/automakers that are lagging in the AV race. Start-ups such as Lvl5, DeepMap and Civil Maps have attracted mapping engineers from Tesla, Apple and Google and these have raised around $40 million in funding. While start-ups such as DeepMap are focusing solely on maps that are meant for fully driverless cars, companies such as TomTom and HERE are focusing on HD maps for advanced driver assistance technologies, which will be helpful prior to the mass adoption of AVs.

Unlike the traditional digital navigation maps, the HD maps need to be updated as regularly as daily in certain regions. For instance, a busy road that goes via pedestrian areas and/or construction site might need a more frequent update on data vis-à-vis a long and uninterrupted freeway. This can possibly translate into a subscription model that could eventually draw in billions for mapping companies. However, the cost involved in creating HD maps is still a big obstacle, as it entails sending a huge volume of data over the cellular network. In addition, there is another challenge that exists in the form of lack of common standard or approach for these HD 3D maps, nor is there any type of sharing of these critical proprietary data. This creates the need for companies to generate new maps for each city that they want to enter. This may delay the process of deployment in certain areas. The complication is further exacerbated by the existence of different driving rules implemented in each city which would entail tweaking the software and therefore translates into redoing the software for each new geography.

On a broader perspective, Televisory believes, the role of mapping will vary based on the level of technological advancement in the AV industry. In the short term, the mapping is going to be crucial for the development of AVs (whether driver assisted or fully autonomous). The process of building these maps to the point of precision, however, is quite tedious and cumbersome and nobody knows who will be the first one to get this right. Further, with the market flooded with several players, this sector is bound to witness a consolidation in the future as companies currently operate with very high fixed costs. Addedly, there is a possibility of a regulatory backlash as there were recent testing crashes associated with Uber and Google’s Waymo during the 1st half of the year (2018). The privacy issues might also come to the fore, making it pertinent for companies to tweak their software amid associated concerns, thereby necessitating government-industry engagements. A couple of years down the line, AVs are likely to depend more on LIDAR sensors or light detection and ranging sensors rather than maps alone. In fact, maps would then play a secondary role in ensuring safety with LIDAR sensors being used for navigation. In the long run, experts expect AVs to be as smart as humans, thereby making them capable enough to handle all existential problems on a daily basis on their own. That, however, will possibly be achieved in several years from now.