Synchronizing Multi-Vehicle Log Data for Autonomous Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing technologies face challenges in generating accurate environment data, particularly ground truth data and map data, due to uncertainties and incomplete tracking of objects in sensor data from autonomous vehicles.
Innovation Solution
The system generates environment data by utilizing time-related log data from multiple autonomous vehicles operating in proximity, synchronizing and interpolating their log data to determine the timing relationships and locations of objects, thereby improving data accuracy and completeness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If sensor data is captured by autonomous vehicles to generate ground truth data and maps, then environment data can be generated for development and mapping, but the data contains uncertainties and incomplete tracking that reduce reliability
Solution Approach 1:
The patent merges sensor data from multiple autonomous vehicles to generate complementary observations. By combining data sources, the system fills gaps in individual vehicle observations and creates more complete ground truth data, directly addressing the incompleteness problem while improving reliability through cross-validation
Solution Approach 2:
The patent introduces a centralized server as an intermediary that receives, synchronizes, and processes log data from multiple vehicles. This intermediary coordinates the data fusion process, manages timing relationships between different vehicles' observations, and generates unified ground truth data that resolves uncertainties in individual perception systems
2Measurement precision
If log data from multiple autonomous vehicles is synchronized and interpolated to improve data accuracy, then ground truth and map data quality enhances, but system complexity increases
Solution Approach 1:
The patent implements feedback mechanisms where the server receives log data from vehicles, processes it to generate ground truth, and can send corrections or refinements back to the vehicles. This iterative feedback loop improves measurement precision by continuously refining object location data while managing complexity through automated processing
Solution Approach 2:
The patent creates copies of sensor data and perception information from multiple vehicle sources. By maintaining replicated data across different vehicles and processing layers, the system can compare observations, identify discrepancies, and generate more accurate ground truth without requiring complex real-time coordination of all original data sources
3Adaptability or versatility
If perception data with degree of uncertainty is used for autonomous vehicle operation, then vehicles can operate in dynamic environments, but the uncertainty makes difficult utilization of the data
Solution Approach 1:
The patent replaces uncertain perception data processing with a systematic data fusion approach using servers and centralized processing. Instead of trying to handle uncertainty in real-time vehicle operations, the system substitutes a batch processing mechanism that synchronizes and validates data across multiple vehicles, making the data more reliable for operational use
Data Source
AI summary
Techniques are disclosed for utilizing synchronized robot observations. The techniques may include receiving first log data associated with a first vehicle, receiving second log data associated with a second vehicle, determining that a first portion of the first log data is time related to a second portion of the second log data, and using the first portion and the second portion for mapping and/or training models.