Plausibility Check Module for Driver Assistance Object Recognition
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing driver assistance systems face challenges in reliably recognizing new vehicle models and traffic signs due to continuous market introductions and legislative changes, leading to outdated sensor data and increased IT security risks, making it difficult to maintain accurate object recognition without user intervention.
Innovation Solution
A plausibility check module is introduced to classify objects using sensor data and reference information from other vehicles or infrastructure, allowing for automated and decentralized updates, with the option to train the classification module using external servers for aggregated learning data, and utilizing two separate classification modules for official recognition and additional learning.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If the classification module uses fixed parameters for object recognition, then the system maintains stability and official approval, but it cannot recognize new vehicle models and traffic signs introduced after manufacturing
Solution Approach 1:
The system divides the classification module into two separate modules: a first classification module with fixed parameters for official approval and stable object recognition, and a second classification module with adaptable parameters for learning new objects. This segmentation allows each module to fulfill its specific function without compromising the other.
Solution Approach 2:
A plausibility check module acts as an intermediary between the two classification modules, comparing results from both modules and determining which result to trust. This mediator enables the system to safely integrate new object recognition capabilities while maintaining reliability for approved objects.
2Adaptability or versatility
If the system continuously updates classification parameters to recognize new objects, then adaptability improves, but IT security risks and system reliability deteriorate
Solution Approach 1:
The system performs preliminary classification using the fixed-parameter first classification module before allowing updates from the second module. This preliminary action ensures that baseline safety and reliability are maintained before any adaptive changes are applied.
Solution Approach 2:
The plausibility check module provides feedback by comparing results from both classification modules and validating whether new object recognitions are plausible before integrating them into the official recognition system, thereby filtering out potential security risks.
3Reliability
If manual updates are required for new object recognition, then system reliability is maintained through human oversight, but productivity and ease of operation decrease
Solution Approach 1:
The second classification module automatically learns and adapts to new objects through continuous data processing and pattern recognition, eliminating the need for manual updates. The system serves itself by autonomously improving its recognition capabilities while the plausibility check ensures reliability.
4Adaptability or versatility
If a single classification module is used, then device complexity is reduced, but the system cannot simultaneously maintain official approval and learn new objects
Solution Approach 1:
The system divides the classification module into two separate classification modules with distinct functions: one for maintaining official approval with fixed parameters, and another for learning new objects with adaptable parameters. This segmentation enables dual functionality while keeping each module's complexity manageable.
Solution Approach 2:
The plausibility check module merges the outputs of both classification modules, intelligently combining their strengths to achieve both official approval compliance and new object recognition capabilities within a unified system architecture.
Data Source
AI summary
A plausibility check module for a vehicle driver assistance system, including at least one sensor for detecting the vehicle surroundings, and Kl-module(s) to classify objects in the surroundings based on the sensor data supplied by the sensor with an internal processing chain established by parameters, the plausibility check module receiving pieces of reference information about objects in the surroundings supplied by other vehicles and/or by an infrastructure, and comparing the pieces of reference information with the classification result by the Kl-module and to initiate at least one measure for a deviation established by the comparison, so that the parameters of the processing chain of the Kl-module are adapted to the effect that the deviation is reduced in comparable situations. Also described are a driver assistance system having the plausibility check module and Kl-module(s), a method for calibrating a sensor for detecting the vehicle surroundings, and an associated computer program.