Industry Trend Analysis - OEM Autonomous Driving Strategies: A Review - AUG 2017


BMI View: The degree to which OEMs intend to develop their autonomous vehicle capabilities is determined by how each company envisions the role that autonomous vehicles will play in their future business models. Furthermore, those companies seeking to speed up the process of development by involving customers in testing , ru n the risk of jeopardising the legal justification of their operations .

This summary of automotive Original Equipment Manufacturers' (OEMs) strategies towards autonomous vehicle development highlights two broad trends, as automakers often envision very different ways in which their autonomous features will be developed and eventually monetised. Particular differences emerge from whether or not carmakers believe full autonomy - also known as, 'Level Four autonomy' where no driver participation or oversight is needed at all - will become a critical part of their business model. Another key difference emerging is whether or not each automakers are opting to use customer-generated information in the research and development process. This is a significant issue given that decisions on customer involvement also bring with them implications for how fast each company can develop more foolproof autonomous capabilities but also how significant the legal challenges faced by automakers will be.

Automotive OEM* Strategies For Autonomous Driving
Company Broad Strategy Latest Updates Notable Technology Partners
*Also includes software company Google X's strategy. Source: BMI
BMW Relying on more collaborative partnership approach to developing technologies for its iNEXT autonomous range. As part of this, it aims to create an 'open platform' for autonomous driving, suggesting its goal of extending the availability of the technology to other automakers and suppliers. Aiming to eventually establish autonomous fleets, particularly for ride-sharing. Recently partnered with Intel and Mobileye to bring a fully autonomous car to production by 2021. Also released the 'every mini is my mini' business concept for future autonomous car sharing. Intel, Mobileye, Baidu, part of consortium of German automakers controlling Nokia HERE mapping service
Daimler Dual strategy of developing fully autonomous cars and semi-autonomous trucks. Fully autonomous car (no steering wheel and pedals) to be production-ready by 2030 but aims to do so through piecemeal introductions of additional semi-autonomous features with each new model generation. Considering leveraging its Car2Go car-sharing service to provide driverless limousine services in the future. Semi-autonomous truck (with highway autopilot) to be ready by 2025. Introduced new 2017 E300 model with Drive Pilot technology including lane change assist, and automated cruise control. Part of consortium of German automakers controlling Nokia HERE mapping service
Ford Motor Aiming to only release fully autonomous vehicle once it is ready for mass market rather than just luxury lines. Relying on partnerships with technology firms and universities. Phased approach already in play with release of self-parking, automatic emergency braking models. Announced it would triple its test fleet of autonomous vehicles to 30 vehicles and invested USD6.6mn in Civil Maps, a tech start-up which generates 3D maps for autonomous cars. Civil Maps, Microsoft Azure, Velodyne, University of Michigan, Massachusetts Institute of Technology and State Farm, Stanford University, Aachen University
General Motors Company Focused on developing autonomous fleets and technologies through large-scale tech acquisitions as compliments to their own autonomous tech development. Co-developing self-driving fleet of taxis with Lyft is a high priority. Announced it is developing a test fleet of electric autonomous taxis by end-2017. Cruise Automation (acquired), Lyft
Google X By far the furthest ahead in terms of research and development in the field with aim to bring self-driving cars to market by 2020. Has stated its intentions to partner with automakers but has also not completely ruled out eventually manufacturing its own autonomous vehicles. Will collaborate with Fiat Chrysler Automobiles in integrating Google's self-driving technology into 100 new 2017 Chrysler Pacifica Hybrid minivans. Fiat Chrysler Automobiles, Bosch, Nvidia, Continental
Honda Motors Aiming to have full autonomy on highways by 2020. Focused on providing affordable autonomous solutions rather than only making tech available in luxury models. Focused on in-house development rather than mergers, acquisitions and partnerships although is considering GM-Honda R&D partnership to include autonomous vehicle tech. In March, Honda released details of semi-autonomous features (highway driving assistance) in its Civic 2016 model year and its Acura models
Hyundai Motors Much more delayed approach with a goal of bringing full autonomy to market between 2025 and 2030 and semi-autonomous features by 2020. Focusing on developing connected car capabilities first. Partnering with Cisco to improve in-car data processing Cisco
Jaguar Land Rover Jaguar is focused on semi-autonomous features rather than full autonomy for next four years as it aims to preserve the driving experience its vehicles offer. Three technology priorities: Recognizing potential obstacles with cameras, automatic emergency braking, and car-to-car communications for obstacle warnings. Announced it will aim to have 100 semi-autonomous vehicles being tested on public roads by 2020. Engineering and Physical Sciences Research Council (ESPSRC)
Mahindra & Mahindra Drawing from expertise of the many various Mahindra Group subsidiaries to develop 'intelligent mobility' products including autonomous electric vehicles. Will particularly leverage Mahindra's strong presence in the Information Technology sector (e.g. Mahindra Tech). Announced in July it will begin its autonomous vehicle program but gave no target release date. Intra-group co-operation
PSA Peugeot Citroen Taking a phase-in approach with semi-autonomous features first being released from 2018 onwards. Stress on drivers always being able to take back control with no intentions stated so far to establish full autonomy without steering wheels and pedals. Focus on vehicle-to-vehicle and vehicle-to-infrastructure capabilities. In April 2016, two Citroen vehicles drove from Paris to Amsterdam in level 3 autonomous mode (driver present but only to take control if system malfunctioned) TRW (on Driver Assist Systems)
Renault-Nissan Alliance Aim to have at least 10 autonomous models on public roads by 2020 but degree of autonomy will vary across the range. Phased approach: 1. autonomous single-lane highway driving (2016); 2. Adding lane-changing capabilities (2018); 3. Navigating intersections and urban areas (2020). Released first semi-autonomous 'ProPilot' features in the new Nissan Serena. Self-driving features can be used for single-lane, highway driving, heavy traffic driving and include automatic emergency braking.
Tesla Motors Uses customer-based 'beta' testing of self-driving features whereby volunteer customers use self-driving features with data then sent back to Tesla. Algorithms and software are then updated and sent back to customers in the form of over-the-air updates. The company saw its first fatal accident while its semi-autonomous "Autopilot" system was in operation, bringing the system under tougher scrutiny of regulators. Mobileye, Nvidia, Bosch
Toyota Motors As part of USD1bn R&D initiative at the Toyota Research Institute in Silicon Valley, the company is pushing into autonomy through focus on artificial intelligence, software development and robotics with particular help from partnerships with elite US universities. Full autonomy is not the key goal, instead opting for a 'guardian angel' approach where car takes over control only where safety is at risk. Aim to have autonomy capable vehicles on road by 2020. Toyota announced its 'guardian angel' strategy for autonomy in June 2016 ( see Broad Strategy for explanation). Stanford University, MIT, University of Michigan, Microsoft Azure
Volkswagen Group Aims to develop in-house a self-driving system by 2020 with 'several billion euros' worth of investment going into R&D. Audi (a VW subsidiary) is part of the Nokia HERE mapping service consortium
Volvo Cars Taking a more collaborative public-private partnership approach with constant involvement from transport authorities, universities and its major supplier Autoliv. No current intentions to create vehicle capable of autonomously driving through urban environments. Only highway driving focus Recently committed to its Drive Me program from 2017 onwards based on putting 100 families in Gothenburg and London into level 3 autonomous vehicles driven under certain road conditions by the company's IntelliSafe Auto Pilot. Autoliv, national and municipal government agencies, Chalmers University of Technology

Automakers Disagree On Full Autonomy's Importance

As the strategies outlined above highlight, automakers remain divided on the topic of how important it is to develop vehicles that can eventually completely drive themselves without any need for a human driver's input or oversight.

For a number of companies, full autonomy serves as a key element of their strategies to diversify into 'mobility solutions' providers rather than just manufacturers of vehicles. As we have previously highlighted, the introduction of fully autonomous vehicles within 10-15 years will provide a strong boost to the affordability and popularity of ride-sharing services that can utilise these vehicles to create an on-demand driverless taxi service ( see 'Autos Megatrends To 2050: Autonomy, Sustainability & Digital Tech Revolutionising Mobility', March 9 2016). At the same time the growth of these services will encourage consumers to avoid car ownership in favour of using sharing services. The combination of these two factors is, therefore, now forcing some automakers to try and pivot towards offering these kinds of services themselves as traditional direct sales channels start to shrink.

Carmakers who are following or at least openly considering this strategy include BMW, Daimler, and General Motors Company (GM). Each of these companies have already either made significant steps into developing their own car-sharing and ride-hailing services or have partnered with major players in the market (see 'GM-Lyft Partnership Represents Industry's Future', January 11). Thus, expanding these services towards a driverless ride-hailing fleet using their own proprietary autonomous vehicles is a natural progression for these companies. We also highlight the likelihood that the strategies of other automakers like Volkswagen and Ford Motor may also veer in this direction given their recent moves into new sharing-based mobility services.

In contrast, firms such as Toyota Motor, PSA Peugeot Citroen, and Jaguar Land Rover have continued to stress the importance of preserving the role of the driver. Toyota's emphasis on autonomous features as the 'guardian angel' of a driver best sums up this group's views. Instead of erasing the driving experience completely these automakers want autonomous features to enhance the driving experience. For them, autonomy should exist for the purpose of enhancing the safety of drivers and passengers by operating only in dangerous situations or taking control during the more monotonous driving environments such as highway driving.

Although this seems only a subtle difference between the likes of BMW and GM against Toyota, PSA and JLR, it actually highlights the different ways in which each company believes its customers will use their vehicles. While GM sees merit in non-ownership and sharing, the likes of Toyota, PSA, and JLR are placing their hopes in continued consumer interest in actually driving and owning vehicles themselves.

Disagreements Over Placing The Consumer At The Front Line

From an operational point of view, differences in how vehicles are being developed have also emerged. The key differences are to do with how new autonomous features are being tested either with or without consumer involvement and each has its various advantages and disadvantages.

For automakers like Tesla Motors and Volvo Cars, testing of automotive features is increasingly going to be conducted by volunteer customers, which will help speed up the development of these vehicles. By having customers use these autonomous features as part of their day-to-day driving, Tesla's 'Autopilot' program and Volvo's 'Drive Me' program will allow each company to more quickly gather data on how autonomous vehicles will be used by actual customers and increase the 'driverless' hours logged by their autonomous fleets more quickly. This will help them refine their algorithms and products much faster than companies such as Google X, Honda Motor and PSA are able to, given their reliance on only a handful of designated testing vehicles.

However, by putting customers in the testing seat, Tesla Motors and Volvo, are placing themselves at a higher risk of incurring legal setbacks to their autonomous driving developments. This has been made most clear with the recently reported fatal crash of a Tesla customer believed to be using the company's 'Autopilot' feature. The death has put Tesla under the scrutiny of consumer groups and safety regulators around the world and has reminded stakeholders that autonomous features are still very much operating in a legal 'grey area'. Thus, this type of customer-led testing can still place automakers in legal trouble, which could threaten to set back overall autonomous programs. In this way, more cautious testing methods like those used by Honda Motor and JLR, may in many ways be more sustainable, despite the fact they involve a much slower research and development process.