Open Standards for Deep Learning Automotive Solutions

On September 19 at AutoSens Brussels, Stephane Strahm of Kalray will be joining other attendees to discuss options for addressing the increasingly complex challenges facing automotive vision system engineers. Pushed by developments in markets such as Advanced Driver Assistance Systems (ADAS) and autonomous vehicles, maintaining component interoperability in increasingly complex vehicle subsystems is proving to be a big obstacle. For example, in the quest to keep up with the rapid growth of the autonomous vehicle industry, technologists need to master intelligent data compute for tactical driving systems. Specifically, there is an intense market need for awareness of environment, decision making, compute capability, algorithms acceleration, image analysis, path planning, and more. Many in the industry are recognizing that these and other needs can be met with open standards.

Khronos’ open standards have already achieved proven success in machine learning, embedded vision, and heterogeneous compute. To shape the future of safety critical APIs for autonomous software solutions, Khronos has established the Khronos Safety Critical Advisory Forum (KSCAF).

With safety critical (SC) open standards, the automotive industry will be able to increase market growth, reduce costs, and accelerate time to market. In addition, new safety critical efforts promise to:

  • Open up possibilities for interoperability;
  • Allow for versatility in the supply chain;
  • Increase adoption of products;
  • Define compatibility;
  • Open up innovation.

Moreover, a collaborative forum like the KSCAF is best poised to aid the automotive industry’s rapidly advancing connected technology as it distributes the intellectual workload across suppliers with a common interface.

Comprised of both internal and external experts, KSCAF develops guidelines and requirements for engineers creating open standard APIs so that those standards can help streamline the product safety certification process. With experts from within Khronos (who span across fields like semiconductors, image processing, and parallel data compute processing) working together with Khronos-invited experts from the medical, automotive, and avionics industries, KSCAF is encouraging familiarity and innovation in the development of SC open standards, which will ultimately bring low-entry costs, efficiency, and flexibility to automotive. KSCAF’s diversified community also allows for a wider range of development experiences to draw from and combines Khronos’ internal experience with real use cases.

Getting Involved with the KSCAF

The SC guidelines and requirements from the KSCAF are always available to the broad community, but getting involved in the forum offers its own benefits. For example, joining the forum can give you confidence that the APIs or implementations developed are compatible with your functional safety demands. The forum’s efforts will also help accelerate the availability of the SC API to implementers, which, in turn, makes it available to you.

To get involved in KSCAF, all experienced practitioners are invited to apply for Advisory Forum membership, at no cost.

Learn more about the KSCAF from Stephane Strahm on September 19 at AutoSens, where he’ll join the panel discussion, “Can We Have Both Safety and Deep Learning-based Decision-making in Autonomous Vehicles” on 9/19 at 1:10 p.m in the Minerva Seminar Room.

You can also see more information and request to join at the Khronos Safety Critical Advisory Forum Landing Page.

Khronos, EGL, glTF, NNEF, OpenVG, OpenVX, OpenXR, SPIR, SPIR-V, SYCL, Vulkan and WebGL are trademarks or registered trademarks of The Khronos Group Inc. OpenCL is a trademark of Apple Inc. and OpenGL and OpenML are registered trademarks and the OpenGL ES and OpenGL SC logos are trademarks of Hewlett Packard Enterprise used under license by Khronos. All other product names, trademarks, and/or company names are used solely for identification and belong to their respective owners.
Authored by

Comments