IEEE Systems, Man and Cybernetics Magazine - January 2020 - 36

self-driving vehicles is to increase road safety and, thus,
reduce the number of accidents worldwide. Safety has
been an intensively discussed topic for automated vehicles
since the projected complexity of these systems reaches
beyond the capabilities of today's traditional assessment
methods [1]. Current validation and verification tools and
safety standards are applied to specific components and
driver-assistance features, assuming that vehicle control is
still overseen by the human driver. These "passive" systems have been tested and verified according to wellestablished pipelines and processes, such as the
automotive-standard V-model and International Organization for Standardization (ISO) 26262 guidelines [2]. Highly
automated systems (self-driving vehicles with capabilities
beyond the current driver-assistance features), on the
other hand, will require a full assessment of their performance and a safe development pipeline to verify their
readiness for production [3]. This need has been generally
recognized by the engineering community, and, as a result,
the IEEE called to action the Global Initiative, which aims
to support safe autonomous-system development via ethical standards (the P7000-standard series) [4].
The way the automotive industry fundamentally
approaches the problem of safety was undoubtedly influenced by technological transitions during the development and test processes in several other industries. One
of those fields is avionics, where the shift from complete

human control to full automation (still supervised by
humans) experienced significant progress during recent
decades. A common feature concerns the instrument-aided landing categories defined by the International Civil
Aviation Organization, where the level of visibility determines the required amount of automation for conducting
landing maneuvers [5]. Contrary to the currently deployed
automotive driver-assistance systems, bad environmental
conditions require or highly recommend the use of automated features that are fail-passive or fail-operational
and supervised by professionally trained human operators. Globally, such training is not feasible for individual
drivers, and the likelihood of one accident per mile
remains several orders of magnitude higher than in the
aviation industry.
To overcome the compatibility challenges of applying
traditional processes to higher levels of autonomy, this
article proposes a new pipeline where automotive safety is
integrated into the technology-development workflow
from the concept level all the way to the testing and validation of the systems. Simulation as a tool for automotive
testing and development is discussed in detail, and some
of the core challenges for its qualification are highlighted.
Safety-Driven Development
Figure 1 presents a concept of the safety-development
workf low that rethinks the traditional automotive

Validation of Requirements Maturity

Law and
Standards

Free-Side Scenarios
(Randomized)
Features

Developer
Group

Functional
Requirements
Functional Safety

Consistency

Synchronize

System/Safety
Group

Verify
Technical
Safety

Fault Injection
Scenarios

Hardware and Software Realization

Verification Scenarios
(Derived From the
Functional Requirements)

Technical
Solutions

Figure 1. The proposed safety-development workflow.
36

IEEE SYSTEMS, MAN, & CYBERNETICS MAGAZINE Janu ar y 2020
IEEE Systems, Man and Cybernetics Magazine - January 2020

Table of Contents for the Digital Edition of IEEE Systems, Man and Cybernetics Magazine - January 2020
