AV Data Offloading With Manifest Reconciliation Under Bandwidth Limits
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in managing data due to limited computing resources, variable network bandwidth, and unpredictable data transfer times, which affect real-time operation, security, and efficiency in data collection and processing.
Innovation Solution
A scalable and flexible AV data management platform that partitions raw data into ingestion objects, applies transformations, and uses manifests for integrity verification, allowing real-time offloading to data centers and servicing stations, with reconciliation to ensure data integrity and compliance with service level agreements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If autonomous vehicles collect and store large amounts of data locally for offline analysis, then data availability for safety and legal operation is improved, but computing resources and storage capacity are exceeded
Solution Approach 1:
The patent segments the data management system into multiple components: local edge computing devices in vehicles, regional data centers, and centralized cloud platforms. Data is partitioned and distributed across these segments, with critical real-time data processed locally and non-critical offline analysis data stored remotely, thereby reducing local storage and computing requirements while maintaining data availability.
Solution Approach 2:
The patent introduces regional data centers as intermediary components between vehicles and centralized cloud platforms. These intermediaries buffer and pre-process data locally, reducing the burden on vehicle computing resources while ensuring data availability for offline analysis. The intermediaries manage data transmission and storage, balancing local and remote data management requirements.
2Productivity
If autonomous vehicles transfer data in real-time to centralized repositories, then data processing efficiency is improved, but network bandwidth requirements increase and data transfer becomes unpredictable
Solution Approach 1:
The patent implements periodic data synchronization where vehicles transfer data to centralized repositories at scheduled intervals rather than continuously in real-time. This periodic action reduces instantaneous network bandwidth requirements while maintaining data processing efficiency by ensuring data is eventually transferred and processed. Critical safety data may be transferred more frequently, while non-critical data uses longer intervals.
Solution Approach 2:
The patent performs preliminary data filtering and prioritization at the vehicle level before transfer. Data is pre-processed to identify critical information that requires immediate transmission versus non-critical data that can be transferred later. This preliminary action reduces the volume of data requiring real-time network transmission while ensuring critical data processing efficiency is maintained.
3Speed
If autonomous vehicles process data locally for immediate maneuvering, then real-time response time is improved, but computing resource consumption increases
Solution Approach 1:
The patent applies local quality by differentiating data processing requirements based on urgency and criticality. Critical safety-related data receives high-priority local processing with dedicated computing resources to ensure immediate response. Non-critical data such as telemetry and analytics is processed with lower priority or deferred to remote systems, optimizing the allocation of computing resources based on the specific quality requirements of different data types.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A data management platform for Autonomous Vehicles (AVs) is provided. An AV can partition raw data into ingestion objects. The AV can transform the ingestion objects and generate associated manifests. The AV can offload first copies of the manifests in real-time to a data center. At a later time, the AV can offload second copies of the manifests to an AV servicing station. The station can upload the second copies to the data center. If the manifests match, then the station can notify the AV that it is safe to erase the manifests and transformed objects from local storage after offloading completes. If the manifests do not match and a Service Level Agreement (SLA) is violated, then the AV can be docked for further diagnosis. If no SLA is applicable, then the error can be annotated and the transformed objects can be discarded.