Autonomous Driving Data Acquisition Using Virtual Model Self-Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for acquiring autonomous driving learning data face challenges in selectively acquiring data from various locations, efficiently using storage space, and excluding malicious data, while also ensuring high-quality data acquisition from user-operated vehicles.
Innovation Solution
An autonomous driving learning data acquiring apparatus equipped with an information acquisition device, a processor, and a storage system that uses a pre-learned artificial neural network (ANN)-based learning model to determine necessary input data for learning recognition logic, and communicates with a server for data transmission and model updates, ensuring efficient data storage and exclusion of malicious data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If all input data is stored without selective acquisition, then storage capacity is insufficient, but data quality and relevance cannot be ensured
Solution Approach 1:
The system performs self-evaluation of data quality by comparing actual sensor inputs against expected sensor outputs generated by a virtual model of the autonomous vehicle. This self-service mechanism automatically identifies and stores only high-quality learning data without external intervention, resolving the contradiction between storing large quantities of data and ensuring data quality.
Solution Approach 2:
The system implements a feedback loop where the virtual model generates expected sensor outputs that are compared with actual sensor inputs. This feedback mechanism enables continuous evaluation of data quality and automatic selection of valuable learning data for storage, balancing data quantity and quality requirements.
2Quantity of substance
If data is selectively acquired based on location and quality, then storage efficiency is improved, but data transmission time and processing complexity increase
Solution Approach 1:
The system performs preliminary evaluation of data quality using the virtual model and sensor comparison mechanism before data transmission. By pre-identifying high-quality learning data and storing it locally, the system avoids transmitting unnecessary data, thereby improving storage efficiency without significantly increasing processing time.
3Measurement precision
If a virtual model and sensor comparison mechanism is implemented, then data quality evaluation is improved, but device complexity and computational load increase
Solution Approach 1:
The system creates a virtual copy (digital twin) of the autonomous vehicle that replicates its physical characteristics and sensor behaviors. This virtual model serves as a reference for evaluating actual sensor data quality, enabling precise evaluation without requiring complex physical testing infrastructure.
Data Source
AI summary
The present disclosure relates to an autonomous driving learning data acquiring apparatus, which selectively acquires learning data of an autonomous vehicle, and a method thereof. According to an embodiment of the present disclosure, an information acquisition device may acquire input data of recognition logic for autonomous driving. A processor may determine whether the acquired input data is necessary for the learning of the recognition logic, through a pre-learned artificial neural network (ANN)-based learning model. A storage may storage storing input data, which is determined to be necessary for the learning of the recognition logic, from among the acquired input data. Through the present disclosure, it is possible to efficiently use a storage space of an autonomous vehicle's data storage device and to effectively acquire high-quality learning data from the autonomous vehicle driven by users.