The Open Source Solution to Autonomous Safety #smartdrivingcar


Safety and, as importantly, the perception of safety could be the pin that pricks the expectations surrounding the autonomous vehicle future. Recognizing the importance of safety to the success of this still nascent industry, autonomous taxi start-up, Voyage, recently placed their testing and reporting procedures in an open source framework. Voyage Co-Founder, Eric Gonzalez explains in the above interview that, at launch, there are four functional blocks to OAS (Open Autonomous Safety) that are now part of a GitHub repository:

  • Scenario Testing
  • Functional Testing
  • Autonomy Assessment
  • Testing toolkit

Oliver Cameron, Voyage Co-Founder and CEO, is excited to see participation and says, “We can’t wait to have all of these contributions from companies from around the world; contribute to build the actual standard in autonomous safety.”

And the Rest of the Story (added 5/23/18):

The video below is the complete interview with Voyage’s Oliver Cameron and Eric Gonzalez and addresses things such as:

  • Other potential functional elements that could be added in the future, such as cybersecurity
  • The use of GitHub as a repository for OAS information
  • Scenario playback from real-world driving situations
  • Key-performance metrics and how they might evolve

The core idea is openness and transparency regarding safety. To paraphrase Princeton’s Dr. Kornhauser’s message in this GovTech article, safety should not be a competitive advantage.

The OAS framework could provide industry, the public and regulators with an opportunity to speak the same language.

Speaking at the SmartDrivingCar Summit, the CIO for the California DMV, Bernard Soriano, explained that the intent of the DMV’s initial reporting requirements for autonomous were to be somewhat open-ended and not too prescriptive. By allowing the autonomous vehicle license holders to submit data they felt important, it has helped educate the DMV in refining their data collection and reporting mechanisms. The open-source foundation of OAS could help further the DMV’s efforts through a collaborative approach with all stakeholders.

[updated 5/23/18]

Attend next year’s 2019 next week’s 2018 SmartDrivingCar Summit to see the rest of the interview, as well as hear from other leaders in the autonomous vehicle space.  ViodiTV was will be there in support of the 2018 SmartDrivingCar Summit, so stay tuned for ViodiTV video coverage of that event.

Author Ken Pyle, Managing Editor

Comments

4 responses to “The Open Source Solution to Autonomous Safety #smartdrivingcar”

  1. Ken Pyle, Managing Editor Avatar

    In researching the DMV database for the above narrative, it was clear that Waymo is providing the most comprehensive reporting and definitions, as evidenced by the 40+ page document found here:

    https://storage.googleapis.com/sdc-prod/v1/safety-report/waymo-safety-report-2017.pdf

    It explains their measurement metrics and how they have extended the 28 core competencies adapted from research by California Partners for Advanced Transportation Technology (PATH) to 47 (page 36). It’s not clear to this author how these directly relate to OAS. Given Waymo’s market leadership, their participation in OAS will be a critical element to OAS’s success.

    1. Ken Pyle Avatar

      This excellent report from the RiSA2S Lab at SJSU compiles the California DMV crasch data (as of September 2017)

      http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0184952#pone-0184952-g013

  2. Ken Pyle, Managing Editor Avatar

    Complementing OAS is the the database of 100,000 real-world videos released by Cal Berkeley and its Deep Drive program. To download these files (collectively, they are huge), click on this link:

    http://bdd-data.berkeley.edu/

    and described in this whitepaper:

    https://arxiv.org/abs/1805.04687

  3. Ken Pyle Avatar

    And this database of annotated images released by FLIR with the idea that thermal sensing will provide another low-cost sensor input to improve the safety of autonomous vehicles.

    “The first of its kind to include annotations for cars, other vehicles, people, bicycles, and dogs, the starter thermal dataset enables developers to begin testing and evolving convolutional neural networks (CNN) with the FLIR Automotive Development Kit (ADK™).”

    https://www.flir.com/news-center/press-releases/flir-releases-starter-thermal-imaging-dataset-for-machine-learning-advanced-driver-assistance-development/

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.