Selective Logging for Vehicle Network Logger Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing amount of data logged by autonomous vehicle (AV) fleet systems poses challenges in terms of storage costs, offload times, and hardware costs, necessitating a reduction in the total size of data logged and stored.
Innovation Solution
A vehicle network logger system (VNLS) that reduces the size of data logged by compressing data, reducing the types of data logged, and reducing the duration for which data is logged, with a focus on selective logging techniques.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If all sensor data is logged continuously, then complete driving data is captured for analysis and debugging, but storage costs and hardware costs increase significantly
Solution Approach 1:
The patent extracts and logs only specific data elements that are relevant to debugging and analysis, rather than logging all sensor data. This selective extraction approach maintains the necessary data completeness for operational purposes while significantly reducing the overall data volume stored, directly resolving the contradiction between reliability and quantity of data.
2Reliability
If all sensor data is logged continuously, then comprehensive data is available for analysis, but offload time increases
Solution Approach 1:
By extracting and logging only essential data elements, the patent reduces the total data volume that needs to be offloaded from the vehicle to cloud storage. This extraction approach maintains data completeness for analysis purposes while significantly decreasing the offload time required, effectively resolving the contradiction between reliability and time loss.
3Duration of action of stationary object
If data retention period is extended, then more historical data is available for analysis, but storage costs increase
Solution Approach 1:
The patent applies selective extraction of data elements, logging only those specific parameters that are most valuable for long-term analysis and debugging. This approach enables extended data retention periods with reduced data volumes, as the extracted subset of critical data maintains analytical value over time without requiring storage of all possible sensor data, thus resolving the contradiction between duration and quantity.
Data Source
AI summary
A VNLS is described and includes stack recorders for recording data from active data sources comprising at least one topic and having associated therewith a logstream comprising a plurality of active data source data elements; edge recorders for recording data from passive data sources and having associated therewith a second logstream comprising a plurality of passive data source data elements; a staging area for storing the first and second logstreams; and a library for storing snapshots associated with events, wherein the snapshots are created when triggers associated with the events are received at the VNLS and wherein the snapshots include links to a first subset of the plurality of data elements; wherein a second subset of the plurality of data elements that are outside a rolling hold time window and that are not in the first subset of the plurality of data elements are periodically deleted from the VNLS.