Driving Assistance Validation for Vulnerable Road User Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current driving assistance systems for autonomous vehicles face challenges in accurately identifying vulnerable road users, leading to potential deactivation of safety features due to incorrect object identification, which can negatively impact road safety.
Innovation Solution
A driving assistance system that includes an object identification module, an analysis module to validate the categorization of vulnerable road users based on movement properties, and a filter algorithm to check the plausibility of identified objects, using environmental data from sensors like LiDAR, radar, and cameras to differentiate between types of road users.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If sensor measurements from various sensor types are used to enable precise environmental perception, then the quantity and diversity of detection data is improved, but the difficulty of ensuring accurate environmental perception increases due to different physical measurement principles
Solution Approach 1:
The patent introduces an intermediary validation process that acts as a mediator between the object identification module and the driving assistance module. This validation process checks the plausibility of identified objects by analyzing movement patterns, serving as a bridge that reconciles the diverse sensor data with accurate environmental perception requirements.
Solution Approach 2:
The patent implements a feedback mechanism where the validation process continuously monitors identified objects and provides feedback to confirm or reject their categorization. This feedback loop ensures that only plausibly identified objects influence driving assistance decisions, improving measurement precision through iterative verification.
2Difficulty of detecting and measuring
If object identification is performed based on sensor data, then the ability to detect road users is improved, but the reliability of identification deteriorates due to incorrect categorizations such as mistaking motorcycles for cyclists
Solution Approach 1:
The patent applies preliminary action by performing validation of object identification before the driving assistance module makes decisions. The system pre-checks the plausibility of identified objects by analyzing their movement patterns, ensuring that only reliably categorized objects influence subsequent driving assistance actions.
Solution Approach 2:
The patent uses dynamics by analyzing the movement patterns of identified objects over time. Instead of relying on static classification, the system continuously monitors changes in position, speed, and acceleration to dynamically validate whether an object's behavior is consistent with its identified category, thereby improving reliability.
3Reliability
If driving assistance is deactivated when vulnerable road users are identified, then road safety is improved, but the service life and usability of the driving assistance system deteriorates due to unnecessary deactivation from incorrect identifications
Solution Approach 1:
The patent performs preliminary validation of object identification before triggering driving assistance deactivation. By pre-checking the plausibility of identified vulnerable road users through movement pattern analysis, the system ensures that deactivation only occurs when identifications are reliable, preventing unnecessary service interruptions.
Solution Approach 2:
The patent implements a feedback mechanism that validates object categorization before influencing driving assistance status. This feedback loop ensures that only accurately identified vulnerable road users trigger deactivation, maintaining both road safety and system usability by preventing false positives from affecting service continuity.
Data Source
AI summary
The present disclosure relates to a driving assistance system for a vehicle, including an object identification module which is configured to carry out an object identification process in a vehicle environment based on environmental data from an environmental sensor system of the vehicle; an analysis module which is configured to validate a categorization of an object as a vulnerable road user, the categorization having been carried out by the object identification module, wherein the analysis module is configured to carry out the validation based on at least one movement property of the object; and a driving assistance module which is configured to carry out driving assistance for automated driving based on the validation.