Vehicle Function Release Workflow for AI Admissibility Checks

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing AI-powered vehicle systems face issues with autonomous decision-making within predefined limits, leading to potential malfunctions, uncontrolled interventions, and homologation-relevant changes, while also limiting the system's potential for functional modifications.

Innovation Solution

A method involving a development device with AI to create and check vehicle function designs using predefined criteria, followed by a remote test device with AI to ensure compliance with additional criteria, ensuring safe and cost-effective implementation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If AI systems operate autonomously within predefined rules and limits, then decision-making efficiency is improved, but the risk of unforeseen interactions and malfunctions increases

Engineering Contradiction:

Improvedecision-making efficiencyVSAvoidsystem reliability

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

A human-in-the-loop verification mechanism is introduced as an intermediary between AI decision-making and system execution. The operator receives explanations for AI decisions and can approve, modify, or reject them before implementation, preventing autonomous malfunctions while preserving AI efficiency.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system implements feedback loops where AI decisions are monitored, evaluated, and used to refine future decisions. Operator corrections and system performance data feed back into the AI model, improving reliability over time while maintaining efficient autonomous operation.

Inventive Principle:

Principle #23Feedback

2Adaptability or versatility

If AI intervention limits are set broadly to allow full potential, then functional adaptability is improved, but uncontrolled interventions may cause homologation-relevant changes

Engineering Contradiction:

Improvefunctional adaptabilityVSAvoidhomologation compliance

Core Design Contradiction:

Adaptability or versatilityVSReliability

Solution Approach 1:

Admissibility criteria and boundaries are predefined before AI operation begins. These pre-established rules constrain AI interventions to acceptable ranges, ensuring homologation compliance while allowing broad functional adaptability within those boundaries.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system dynamically adjusts AI intervention levels based on real-time conditions and predefined thresholds. AI autonomy is flexible rather than fixed, expanding or contracting based on operational context while maintaining compliance through operator oversight and configurable limits.

Inventive Principle:

Principle #15Dynamics

3Reliability

If AI intervention limits are set restrictively to ensure safety, then system reliability is improved, but the AI system's potential for modifications is reduced

Engineering Contradiction:

Improvesystem safetyVSAvoidfunctional modification potential

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

Safety limits are dynamic rather than static, allowing the system to operate more autonomously when conditions are favorable and requiring more oversight when risks increase. This maintains high safety standards while preserving AI adaptability across different operational contexts.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

Admissibility criteria are segmented into multiple layers (mandatory constraints, preferred guidelines, optional optimizations). This hierarchical structure allows restrictive safety rules to coexist with broader adaptability potential, as not all criteria carry the same weight or applicability.

Inventive Principle:

Principle #1Segmentation

4Speed

If autonomous AI systems make decisions without human awareness, then operational speed is improved, but operator awareness and malfunction detection are reduced

Engineering Contradiction:

Improveoperational speedVSAvoidoperator awareness

Core Design Contradiction:

SpeedVSLoss of information

Solution Approach 1:

The AI system provides continuous feedback to the operator about its decisions, reasoning, and confidence levels. This maintains operator awareness and enables rapid detection of malfunctions while preserving the speed benefits of autonomous decision-making for routine operations.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

An explanation layer acts as an intermediary between autonomous AI processing and human operator awareness. The AI generates human-readable explanations for its decisions, allowing operators to understand and monitor system behavior without slowing down operational execution.

Inventive Principle:

Principle #24Intermediary (Mediator)

Data Source

PatentEP4113348B1Method for developing vehicle functions and motor vehicle

Publication Date: 2025.12.31 VOLKSWAGEN AG

Figure 1
Figure 2
Figure 3

AI summary

The invention relates to a method for developing a vehicle function for implementation by means of hardware (1) of a motor vehicle (2). The method comprises the following steps: - identifying a vehicle function to be developed, - creating a design of the vehicle function, - performing a first admissibility check of the design of the vehicle function, - releasing the design of the vehicle function if the first admissibility criteria are met, - transmitting the design of the vehicle function, - performing a second admissibility check of the design of the vehicle function, - releasing the design of the vehicle function if the second admissibility criteria are met, - transmitting the design of the vehicle function to the motor vehicle (2), and - implementing the vehicle function in the motor vehicle (2). The invention further relates to a motor vehicle (2).