Autonomous vehicles must be able to prove safety by design
| New document, written based on RSS model
| Integration of expertise from automotive manufacturers, suppliers, and professional engineers

To what extent should autonomous vehicles be allowed discretion for safe driving?

Intel announced on the 4th the 'Safety First For Automoated Driving' framework for the design, development, certification, and validation of safe autonomous vehicles in collaboration with 10 automotive and autonomous driving technology companies.

This document is based on Intel's Responsibility-Sensitive Safety (RSS) model for safe autonomous driving decision-making.

To implement a high level of autonomous driving on public roads, it is important to develop autonomous vehicles that can prove their design safety.

The Safety First Framework for Autonomous Driving integrates the expertise of leading global automotive manufacturers, suppliers, and technology providers to provide comprehensive guidelines for the development of autonomous vehicles that are safe by design.

The basic framework of this document consists of 12 principles and the steps required to implement them.

Each principle is established according to a series of functions that a safe autonomous vehicle must support. There are, and safety elements for implementing these functions are introduced.

Details regarding the principles can be found starting from page 5 of the document.

The Intel RSS model was highlighted in the Drive Planning Element, which supports the function of creating a 'collision-free legal driving plan'.

This element understands, predicts, and manages norms regarding autonomous driving, and implements the highest principle of acting safely as a means to comply with road rules.

The vision of reducing traffic accident fatalities through autonomous driving technology presents a broad approach across the industry. The Safety First Framework for Autonomous Driving integrates only the safest approaches among these various methods.

RSS was proposed in 2017 as a technology-neutral starting point to allow the industry to coordinate on what it means for autonomous vehicles to drive safely. RSS is derived from transparent and verifiable mathematical formulas that formalize the human concept of common sense to provide a safety range for the decision-making functions of autonomous vehicles.

The Safety First Framework for autonomous driving accelerates the global adoption momentum of RSS, including recent support from Baidu, Valeo, and China ITS.

Jack Weast, Intel’s chief engineer and Vice President of Autonomous Driving Standards at Mobileye, stated, “Industry collaboration on the safety of autonomous vehicles is key to realizing a safe and responsible future for autonomous driving.”

He added, “Intel is very pleased to contribute to the project of establishing a framework for the introduction of design-safe autonomous vehicles,” and stated, “Intel looks forward to collaborating with new industry partners on the RSS model as well as the broader framework.”