Vehicle Pre-filtering Circuit for Data Volume Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The vast amounts of data collected by vehicles, such as autonomous vehicles, pose significant challenges in storage and communication resources, as well as processing latency, due to the need to transfer large datasets to the cloud for analysis, which is costly and resource-intensive.
Innovation Solution
Implementing a pre-filtering circuit in vehicles that determines and transmits only the necessary data points for extrapolation or interpolation functions, reducing the amount of data sent to the cloud or network edge by analyzing and correlating anomalous data and using scenario patterns to identify required data subsets, thereby simplifying in-vehicle communications and reducing latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If all collected vehicle data is transmitted to the cloud for processing, then complete data analysis is achieved, but storage and communication resource demands increase significantly
Solution Approach 1:
The system performs preliminary filtering of vehicle data at the edge device before transmission to the cloud. By pre-processing the data and identifying only relevant data points that meet specific criteria, the system reduces the volume of data transmitted while ensuring that complete analysis is performed on the essential subset of data.
Solution Approach 2:
The system extracts only the necessary data points from the collected vehicle data based on predefined criteria and scenario patterns. This extraction process removes redundant information while retaining the essential data needed for comprehensive analysis, thereby reducing transmission volume without sacrificing analytical completeness.
2Manufacturing precision
If large datasets are transmitted to the cloud, then thorough AI/ML model training is enabled, but processing latency increases
Solution Approach 1:
The system performs preliminary filtering and selection of relevant data points at the edge device before cloud transmission. This pre-processing step ensures that only high-quality, relevant data is transmitted, enabling efficient model training with reduced latency while maintaining training accuracy.
Solution Approach 2:
The system transmits a partial set of data points that are specifically selected to be sufficient for accurate model training. By transmitting only the necessary subset of data rather than complete datasets, the system achieves adequate training accuracy with significantly reduced processing time.
3Reliability
If comprehensive vehicle data is collected and stored, then complete scenario analysis is achieved, but storage resource requirements become prohibitive
Solution Approach 1:
The system extracts only the essential data points needed for complete scenario analysis from the collected vehicle data. By identifying and retaining only the critical data elements that contribute to scenario understanding, the system achieves reliable analysis with minimal storage requirements.
Solution Approach 2:
The system performs preliminary filtering and categorization of vehicle data at the edge device, organizing data into scenario patterns and identifying only the relevant data points for each scenario. This pre-processing enables complete scenario analysis while significantly reducing the storage capacity needed.
Data Source
AI summary
Systems and methods are provided for pre-filtering vehicle-related data obtained from vehicle sensors, V2X communications with roadside infrastructure or vehicles, and/or third-party information sources. The amount of data received from such data sources can be massive. The systems and methods pre-filter the data at the vehicle prior to transmission to an artificial intelligence or machine learning system for analysis so that the amount of data transmitted can be reduced, easing the demand on communication and data processing resources. Moreover, the speed at which the transmitted data can be analyzed is increased relative to conventional systems that rely on characterizing scenarios, training models, predicting events, etc. using as much information as can be collected.