The Background and Motivations Behind the Software-Defined Vehicle (SDV)

The Software-Defined Vehicle (SDV) is far more than a vehicle equipped with extensive software. It represents a fundamental shift in how vehicles are developed, how they function, and how they evolve over time. Software becomes the central design element, while hardware, software architecture, and development processes are systematically aligned.

The motivation behind this concept extends well beyond infotainment and digital convenience features. Software-based control and decision-making methods enable many modern functions in powertrains, chassis systems, and vehicle safety. They also provide an essential foundation for automated, autonomous, and connected driving.

However, the true potential of the SDV lies not only in new features. What matters is the ability to develop complex systems systematically, ensure their safety, and continuously evolve them throughout their entire lifecycle.

This article explores the key technical and organizational factors behind the SDV and examines the roles played by functionality, artificial intelligence, development processes, system architecture, and connectivity.

The primary motivation behind the SDV is to enable more capable vehicle functions. This applies not only to infotainment but, in particular, to powertrains, chassis systems, active safety and driver assistance systems, as well as automated and connected driving.

Digital control algorithms make it possible to process and coordinate sensor information, vehicle states, and control inputs in real time. This enables more precise vehicle dynamics control, better utilization of stability margins, and more targeted adaptation of safety systems to different driving situations.

Artificial intelligence techniques are also increasingly being used for complex functions. They expand the ability to identify patterns in data and solve demanding decision-making tasks.

Automated and connected driving illustrate this development particularly well: their functionality emerges from the interaction of sensors, software, control engineering, and digital communication.

A key advantage of software-defined vehicles is that many functions can evolve through software changes without necessarily requiring mechanical or electronic components to be replaced. This requires an appropriate system architecture, together with suitable update and release processes.

New functions can initially be deployed within a limited operational scope and then expanded incrementally. In automated driving, for example, a function might first be deployed under clearly defined operating conditions. Its operational limits can be extended only once the required performance and safety have been demonstrated.

The Operational Design Domain (ODD) describes the operating conditions for which an automated driving function has been designed. Expanding this domain requires the additional conditions to be systematically analyzed, developed for, and validated.

Data collected during real-world operation can also contribute to improving future software versions. The key is to incorporate the insights gained from these data into a controlled development, testing, and release process.

Artificial intelligence expands the possibilities for implementing complex vehicle functions. In particular, machine learning methods can identify patterns and relationships in data that may be difficult to capture using conventional, entirely rule-based approaches.

However, these capabilities also introduce additional development and safety assurance requirements. A model's performance must be evaluated using appropriate data and representative scenarios. This assessment must cover not only typical situations but also rare events, disturbances, and the limits of the intended operating domain.

Depending on the method used, learning takes place during a training phase, or adaptation is performed through a separately controlled process. Changes to models and algorithms must be documented in a traceable manner, evaluated, and appropriately approved before deployment.

The development of AI-based vehicle functions therefore requires a closed-loop process encompassing design, training or implementation, validation, release, and systematic further development.

The potential of an SDV does not arise from modern software alone. The processes used to specify, develop, integrate, test, and release vehicle functions are equally important.

Requirements-driven development ensures that technical solutions are derived from the desired system functions and their associated requirements. Particularly for complex driving functions, these requirements must be examined early for consistency, completeness, and feasibility.

The development model is also changing. Rather than evolving software exclusively within individual vehicle projects, the goal is to enable continuous improvement throughout the entire product lifecycle. New functions and corrections are developed, integrated, and verified incrementally.

Standardized and largely automated processes help reduce development effort, identify errors earlier, and integrate changes into the overall system in a controlled manner.

Ideally, this approach to further development can scale across multiple vehicle types.

The continuous evolution of an SDV requires changes to be reliably evaluated for their effects. This includes automated testing at component, integration, and system levels, as well as verification of the specified vehicle functions.

For safety-critical functions, particular attention must be paid to potential interactions between individual components. A change to one software component can affect the behavior of the overall system, even when the change itself is technically correct.

Scenario-based testing and simulation make it possible to investigate defined driving situations reproducibly and systematically compare system responses against requirements. Depending on the function and the intended evidence, additional testing and assessment methods are also necessary.

Automated testing and integration processes are therefore an essential prerequisite for short development cycles. However, they do not replace the necessary safety assessment and approval of each relevant change.

An appropriate system architecture is a key prerequisite for the economic success of the SDV. It must enable software functions to be developed modularly, reused, and deployed across different vehicle models.

The architecture should therefore be based consistently on functional requirements and the logical relationships within the overall system. Hardware components remain important technical constraints, but they should not unnecessarily restrict the structure of the software.

Separating functionality from technical implementation can increase software reusability, reduce dependencies, and simplify further development. This requires clearly defined interfaces, end-to-end integration, and close coordination between system architecture, software development, and the organizational units involved.

Connectivity expands the capabilities of software-defined vehicles. It provides the technical foundation for delivering over-the-air (OTA) software updates, analyzing vehicle data under appropriate conditions, and implementing connected functions.

This enables vehicles to evolve beyond their original delivery. However, doing so requires suitable technical and organizational safeguards, particularly for secure update delivery, protection against unauthorized access, and responsible data handling.

Applying agile software engineering methods is not merely a matter of how software is developed. It also requires and promotes a flexible organizational structure. Cross-functional teams, short feedback cycles, and close collaboration across disciplines help prevent organizational silos from becoming entrenched in rigid responsibilities and communication channels.

This can provide an important safeguard against the effects of Conway's Law, which states that a system's architecture tends to reflect the communication structures of the organization that develops it. This relationship is particularly significant for the Software-Defined Vehicle, whose functionality emerges from the complex interaction of numerous hardware and software components.

Agile organizational and development structures support an architecture that reflects the functional relationships of the overall system rather than technically reproducing existing departmental boundaries. However, this requires agility to be embedded not only within individual teams but also consistently across decision-making, responsibilities, and collaboration structures.

The Software-Defined Vehicle represents a fundamental shift in vehicle development. Its potential arises from the combination of capable software, suitable system architectures, data-driven methods, and integrated development and safety assurance processes.

New features alone are not enough. What matters is the ability to evolve complex vehicle systems in a controlled manner, verify changes in a traceable way, and maintain the required level of safety throughout the entire lifecycle.

The SDV is therefore not merely a technological development; it also represents a transformation in how vehicles are conceived, developed, validated, and operated.

ANDATA is pioneer and cofounder of the SDV in the domain of vehicle safety. Already in the mid-2000s the underlying principles have been formulated and declined. For the implementation ANDATA has operated a joint venture together with the Volkswagen group for more than 15 years, dedicated to the topic.

The underlying solutions are offered by ANDATA in form of ScenarioManagement-as-a-Service, amongst others. Within VERONET, ANDATA extends the topic generally to the holistic traffic system at all.