Access Official Technical PDF

Share Summary:LinkedInEmail

Automotive functional safety is no longer a concern that can be deferred until an electronic control unit or vehicle system is ready for final integration. As vehicles rely on more sophisticated software, processors, sensors, and connected features, safety-related decisions need to reach into semiconductor design and software development. A shift-left approach brings those considerations forward, giving engineering teams more opportunity to identify and address risks throughout the development process.

This matters because electronic failures in a vehicle have consequences that differ sharply from a software fault on a desktop computer. A failure while a vehicle is moving can put occupants and others at risk. For automotive engineers, system integrators, and program leaders, the challenge is to coordinate functional safety, software quality, verification, and cybersecurity across increasingly complex systems.

Automotive functional safety starts earlier in the design process

ISO 26262 defines requirements for functional safety in road-vehicle electrical and electronic systems, including Automotive Safety Integrity Levels (ASILs). ASIL-D is the highest level described in the source material and can be particularly demanding to achieve. Safety considerations may apply across a system, including its constituent electronic components, rather than only at the vehicle or ECU level.

That makes the system-on-chip (SoC) an important point of attention. Integrating safety-related design and verification considerations at the chip level can help teams begin addressing them before hardware reaches the broader system-development stage. The objective is not to treat chip design as a substitute for system-level safety work, but to make functional safety part of development across multiple levels.

Virtual prototypes can support this earlier work. These models of hardware under development allow software development to begin before physical hardware is available. Earlier access to a representative platform can help teams start software work and integration sooner, while reducing dependence on physical systems as the sole test environment.

SoC complexity raises the bar for safety and verification

Automotive architectures are evolving from distributed ECUs toward more integrated domain controllers, while autonomous-driving, advanced driver-assistance systems (ADAS), and infotainment increase the functionality concentrated in vehicle electronics. As SoCs become more complex, their design must account for safety requirements alongside processing needs and implementation constraints.

Verification is a substantial part of that effort. Teams need to address ISO 26262-related requirements for both IP and SoCs, while managing schedules and engineering resources. An integrated process can connect failure analysis, functional safety verification, and the creation of required work products, rather than leaving these activities as isolated tasks.

Different verification methods can contribute at different points in the flow. Fault campaigns may use simulation, formal methods, and emulation, with concurrent engines helping teams apply suitable approaches to complex designs. Static analysis of soft-error propagation, at RTL or gate level, can help assess how bit flips may travel through a design and identify areas for attention. The resulting analysis can inform measures such as the Single Point Fault Metric (SPFM).

Built-in self-test as an operational safety measure

Built-in self-test (BIST) provides another tool for safety-critical designs. Logic BIST can test logic at power-up and periodically during functional operation. This supports the detection of faults during the device’s lifecycle, not only during development. Safety management architecture can also coordinate safety bring-up, functional safety testing, memory and logic BIST control, monitoring, and safety escalations across an SoC.

ADAS processing combines performance with safety constraints

Camera-based perception illustrates the competing demands placed on automotive processors. Deep-learning applications have progressed from tasks such as pedestrian detection and collision avoidance toward scene segmentation, in which each pixel in a high-resolution video stream must be interpreted. Cameras and image sensors provide visual input, but processors must analyze that information quickly enough to support the application.

Embedded vision platforms therefore need to balance performance with power and area constraints. Tool support is also important: classical computer-vision development may use standards and languages such as OpenCV, OpenVX, OpenCL, and C/C++, while deep-learning applications rely on frameworks such as TensorFlow and Caffe, as well as model exchange approaches such as ONNX. For automotive teams, the design question is how to map these workloads efficiently while accounting for applicable functional safety requirements, including ASIL B and ASIL D contexts.

Software quality and cybersecurity belong in the same conversation

Functional safety is only one part of dependable automotive development. Faulty software can affect vehicle operation, while connectivity and expanding digital services create cybersecurity concerns. Security-by-design aims to identify risks early enough for suitable measures to be incorporated. Monitoring can then help organizations assess whether protections are effective and reveal shortcomings that may need attention.

Software development practices also influence how quickly teams can build and test increasingly complex systems. Continuous integration and continuous delivery (CI/CD), along with DevOps practices, can support faster development cycles, but adopting them in automotive programs requires attention to software quality and security. Virtual test environments can help reduce dependence on physical hardware for development, integration, and testing, particularly when physical systems are a bottleneck.

These approaches are complementary: virtual platforms can bring software work forward, automated workflows can support iterative development, and verification and security activities can be incorporated into the process rather than postponed. Their value depends on fitting them to automotive constraints and maintaining appropriate safety and quality controls.

Plan safety and security across the development lifecycle

For engineering and operations leaders, the practical implication is to treat safety, software quality, and cybersecurity as connected development responsibilities. Begin by identifying relevant requirements at the system, SoC, and software levels. Use early models where they can enable software work before hardware is ready, and select verification methods that address the design’s risks and complexity. Include operational test mechanisms such as logic BIST where appropriate, and establish processes for assessing security measures as systems become connected.

Moving these activities earlier does not eliminate the need for system-level integration or validation. It gives teams a structured way to address important concerns sooner, coordinate across hardware and software disciplines, and manage the demands of increasingly capable automotive electronics.

Access Official Technical PDF

About The Author