Ring Memory Data Validation for Automated Driving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
There is no generally accepted approach for validating driver assistance systems and partially automated vehicle operations due to error-prone object detection algorithms, which can lead to incorrect decisions in situations requiring protective measures, and existing methods for data validation are time-consuming and costly.
Innovation Solution
A method combining the 'Trojan Horse' and 'Open-Loop' approaches, utilizing a ring memory to store relevant data from series-production vehicles for offline analysis, allowing for rapid validation of driver assistance functions by detecting significant events and updating scenario catalogs with detected data probabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If driving tests are conducted to verify target figures for error rates, then validation accuracy is improved, but the time required extends over several hundred years
Solution Approach 1:
The patent applies preliminary action by pre-collecting and storing sensor data from various driving situations in a database before validation is needed. This allows the validation process to use pre-prepared data instead of conducting extremely long driving tests, thus achieving accurate validation results in a practical timeframe.
Solution Approach 2:
The patent uses copying by creating a virtual representation of driving situations through stored sensor data and scenario catalogs. Instead of physically conducting hundreds of years of driving tests, the system copies relevant driving scenarios from the database and uses them for validation, dramatically reducing the required time while maintaining accuracy.
2Loss of information
If all sensor data is stored for validation analysis, then data completeness is improved, but storage requirements and processing complexity increase significantly
Solution Approach 1:
The patent applies local quality by differentiating between types of data and applying different storage strategies. Frequently occurring or critical driving situations are stored in detail, while less critical data is stored with reduced detail or not at all. This selective approach maintains data completeness for important scenarios while reducing overall storage requirements and processing complexity.
Solution Approach 2:
The patent extracts only the essential and relevant features from raw sensor data for storage and validation purposes. By taking out only the critical information needed for validation (such as key event markers, scenario-relevant parameters) rather than storing all raw data, the system achieves adequate data completeness for validation while significantly reducing storage and processing complexity.
3Productivity
If passive Trojan Horse functions run concurrently in series-production vehicles, then data collection speed is improved, but data assessment difficulty increases due to inactive function state
Solution Approach 1:
The patent uses an intermediary approach by introducing a separate evaluation system that processes the data collected from passive Trojan Horse functions. This intermediary evaluation system simulates what the activated function would have done, bridging the gap between the passive data collection and the need to assess active function performance. This allows rapid data collection while maintaining accurate assessment capability.
Solution Approach 2:
The patent applies copying by creating a virtual model of the activated function's behavior based on the passive data collection. The evaluation system copies the logic and decision-making processes of the activated function and applies them to the collected data, enabling accurate assessment of what the function would have achieved had it been active, thus resolving the assessment difficulty while maintaining rapid data collection.
Data Source
AI summary
A method for detecting data of a vehicle operated at least partially in an automated manner, including at least one sensor, in particular, a surroundings sensor, in particular, a video sensor, a radar sensor, a LIDAR sensor, an ultrasonic sensor, an infrared sensor, and/or a GNSS sensor, in particular, for receiving a GPS signal, a GLONASS signal or a Galileo signal, and/or a vehicle sensor, in particular, for engine control, for activating occupant protection means, for activating assistance functions and/or convenience functions, in particular, an acceleration sensor, a rotation rate sensor, a pressure sensor, and at least one memory, in particular, a ring memory, including detecting data of the at least one sensor; saving the data in the memory, in particular, in the ring memory; detecting an event; initially storing the content of the memory, in particular, of the ring memory, at the point in time of the detected event.