The information & technology industry has been open sourcing intellectual property, databases, platforms, and APIs for a while now. More recently, this trend has gained momentum in the automotive industry. Players in the automotive industry are adopting the open sourcing culture and strategies to keep pace with a rapidly evolving Connected, Autonomous, Shared, and Electric (C.A.S.E) ecosystem.

Disruptors like Tesla open sourced all their patents on electric vehicle (EV) technology in 2014, citing their commitment to reducing their carbon footprint and encouraging the automotive industry to benefit from a common and continuously developing technology platform that would accelerate vehicle electrification initiatives.

The open sourcing of Tesla’s suite of patents was followed by a traditional vehicle manufacturer like Toyota that released about 6,000 patents related to its hydrogen fuel cell technology in 2015, followed by around 24,000 patents related to hybrid car technology earlier this year. This step was aimed at keeping hybrid vehicles relevant as the industry progressively shifts toward all-electric vehicles.

This has been the year for autonomous driving datasets. Alphabet’s autonomous driving unit, Waymo, recently surprised the industry by releasing its Waymo Open Dataset, following Argo AI’s release of its Argoverse dataset and Lyft’s Level 5 dataset.

Waymo’s Open Dataset

Like other autonomous driving technology companies, Waymo too has built its Open Dataset on a foundation of rigorous testing. Its test vehicles fitted with vision and proximity sensors logged over five million miles, collecting traffic, road and environment data on the public roads of 25+ cities in the U.S., including San Francisco, Phoenix, Kirkland, Washington, and Chandler.

Waymo’s Open Dataset has been compiled in a diverse range of environments (urban and suburban), light conditions (day, night, dawn and dusk), weather patterns (sunshine, cloudy, overcast and rainy), driving scenarios (construction zones), and road users (pedestrians and cyclists). It also includes 3,000 driving scenes, 600,000 frames, and approximately 25 million 3D and 22 million 2D bounding boxes collected from its autonomous vehicles that are packaged with a sensor set that includes five LiDARs, five cameras, and a number of radars sensors.

Although Waymo has completed millions of test driving miles and the Waymo Open Dataset encompasses a wide range of scenarios, the miles and data captured are limited to cities and suburban areas in the U.S.

Waymo’s potential approach & strategies

With the release of its Open Dataset, Waymo could be looking at five main business strategies:

1. Accelerate development: To promote innovation in machine learning and solve complex problems, Waymo has provided the research community with access to its large dataset and has opened a community page to share the trained algorithms and findings that are being built using its Open Dataset. The company also plans to host challenges aimed at solving specific problems, thereby attracting and identifying highly talented problem solvers from the community.

2. Enrich the dataset: By open sourcing its dataset, Waymo is encouraging other developers in the industry to share their perception datasets on Waymo’s Open Dataset community page, thereby enhancing the size and variety of the larger data set. Perception datasets shared by other developers will not only provide Waymo with access to geographical and traffic data, it will also enable Waymo to develop high definition (HD) maps of regions not covered by its test driving vehicles. The process of gathering information on other regions picks up speed when other developers, presumably having datasets for regions in which Waymo is not active, join the bandwagon.

3. Advance technology through collaboration: With its Open Dataset, Waymo is demonstrating its technological advancement and capabilities to the industry. It is, thereby, indirectly highlighting the need for collaboration within the autonomous driving ecosystem. Forging business collaborations with OEMs and other players from various parts of the world has the potential to empower Waymo to expand its reach into hitherto untapped markets.

4. Promote a common platform: While most large OEMs are developing autonomous driving software in-house, they are also keen on adopting a common platform. Waymo’s Open Dataset has the potential to assist Waymo to develop an open autonomous driving platform (potentially replicate android platform model of mobile phones) that is compliant with regional vehicle regulations, communication protocols, smart city networks, hosts third party applications, and that can be integrated with any production vehicle and service providers’ interface.

5. Support innovative business models: Every open sourcing platform, although free, is always accompanied by an interesting business model. In Waymo’s case, this
could take the form of:

  • Additional services and features offered over and above the open source datasets will need to be purchased, thereby creating new revenue streams.
  • Waymo will leverage its hardware and software suites with partners that may be interested in the dataset or vice versa.
  • Individually, the datasets, hardware, and software will not be as powerful as they will when offered together as a comprehensive and effective solution

Like Google and Android, Alphabet’s Waymo has the potential to be at the epicentre of the autonomous driving ecosystem while OEMs, service providers, aggregators, and other businesses will collaborate to add and extract value out of it. Open sourcing Waymo’s Open Database could represent the first step towards cutting across boundaries to develop an exhaustive database, multi region HD map, fostering collaboration in the ecosystem, thus leading the industry towards accelerated development and deployment of autonomous driving.

Have Questions?

We have Answers.

Schedule a Complimentary Growth Strategy
Dialogue with an Industry Expert



Powering Clients to a Future Shaped by Growth


Share This