Vehicle Sensor Data Familiarity Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle systems generate large amounts of data indiscriminately, leading to inefficient data handling and storage costs, and struggle to identify and interpret significant data relevant for vehicle detection systems, resulting in unnecessary storage and processing of useless data.
Innovation Solution
A method and system that utilize a control unit to receive and process vehicle sensor data, run scene detection operations using a trained vehicle data set, generate vector representations, and identify significant scenario data by determining target object attributes and familiarity, thereby isolating and reporting only relevant data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If vehicle systems collect and store large amounts of data indiscriminately, then the system can potentially capture all relevant information for detection, but the storage cost and data processing burden increase significantly
Solution Approach 1:
The patent extracts only the significant and relevant data from the vast amount of sensor data generated by vehicle systems. By using trained data sets to identify and extract only the meaningful patterns and anomalies, the system avoids storing and processing unnecessary data, thus reducing storage volume while maintaining detection reliability.
Solution Approach 2:
The patent applies local quality by treating different portions of data differently based on their significance. Rather than uniformly processing all data, the system identifies specific regions or types of data that are most relevant to detection tasks and focuses computational resources on those areas, improving efficiency while maintaining accuracy.
2Loss of information
If vehicle systems process and store all generated data, then comprehensive analysis is possible, but the system complexity and processing requirements increase
Solution Approach 1:
The patent performs preliminary action by pre-training data sets with relevant information before actual detection tasks. This pre-processing creates a foundation of knowledge that enables the system to quickly identify significant data during operation, reducing the complexity of real-time processing while maintaining information completeness.
Solution Approach 2:
The trained data set acts as an intermediary between raw sensor data and detection decisions. This intermediary layer filters and interprets data, translating complex sensor inputs into meaningful detections without requiring the main system to directly process all raw data, thus reducing processing complexity.
3Adaptability or versatility
If the system stores all sensor data for future analysis, then retrospective evaluation is enabled, but storage costs and maintenance requirements increase
Solution Approach 1:
The patent extracts only the significant data that is likely to be useful for future analysis and retraining. By identifying and storing only meaningful patterns, anomalies, and edge cases rather than all sensor data, the system maintains data reusability for adaptability while significantly reducing storage requirements and associated costs.
Data Source
AI summary
The present disclosure relates to systems, devices and methods for identifying objects and scenarios that have not been trained or are unidentifiable to vehicle perception sensors or vehicle assistive driving systems. Embodiments are directed to using a trained vehicle data set to identify target objects in vehicle sensor data. In one embodiment, a process is provided that includes running a scene detection operation on vehicle to derive a vector of target object attributes of the vehicle sensor data and generating a vector representation for the scene detection operation and the attributes of the vehicle sensor data. The vector representation compared to a familiarity vector to represent effectiveness of the scene detection operation. In addition, the vector representation can be scored to identify one or more target objects or significant scenarios, including unidentifiable objects and/or driving scenes, scenarios for reporting.