Self-Learned Relevancy Metrics for Vehicle Perception Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current assisted and autonomous driving systems face inefficiencies in determining driving decisions due to the complexity and resource-intensive processing of environmental information, necessitating a more efficient method for processing relevant data.
Innovation Solution
Implementing self-learning methods, such as semi-supervised and self-supervised training, to determine relevancy metrics for perception-related applications, allowing vehicles to ignore irrelevant information and optimize resource allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If complex environmental information is processed to determine driving decisions, then the reliability of driving decisions is improved, but the resource consumption and processing time increase significantly
Solution Approach 1:
The patent segments environmental information into relevant and irrelevant components using sensor data filtering and feature extraction. The system divides the complex environment into discrete detectable parameters (object distance, relative velocity, acceleration) that are independently processed, allowing the system to focus computational resources only on relevant features while discarding irrelevant information.
Solution Approach 2:
The patent extracts essential features from complex environmental data by identifying and isolating key parameters such as object distance, relative velocity, and acceleration. This extraction process removes irrelevant information before processing, reducing the data volume that requires intensive computation while maintaining the reliability needed for safe driving decisions.
2Measurement precision
If comprehensive environmental information is processed, then the accuracy of driving decisions is improved, but the processing time increases
Solution Approach 1:
The patent performs preliminary processing of environmental information by pre-defining relevance criteria and filtering mechanisms. Before the main decision-making process, the system pre-processes sensor data to identify and extract only those features that meet predetermined relevance thresholds, thereby reducing the computational burden during time-critical decision moments while maintaining accuracy.
3Loss of information
If all sensor data is processed, then the completeness of environmental perception is improved, but the resource allocation efficiency decreases
Solution Approach 1:
The patent applies local quality by assigning different processing priorities and resource allocations to different regions or aspects of environmental data based on their relevance. Critical parameters such as objects in the vehicle's path receive higher processing priority and more computational resources, while less critical information receives reduced processing, thereby optimizing resource allocation while maintaining perceptual completeness.
Data Source
AI summary
A method that is computer implemented for self-learning of relevancy metrics for perception related applications, the method comprising: receiving a training dataset that comprises (i) scenario information regarding a scenario faced by a vehicle and involving a road user, and (ii) behavior information indicative of a response of the vehicle to a presence of the road user; determining, based on the scenario information and the behavior information, a relevancy metric providing an indication of an impact of the road user on a driving of the vehicle; and training, using self-learning, a machine learning process to infer the relevancy metric according to the scenario.