HD Map Updating Across Heterogeneous Vehicle Sensor Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The accuracy of high-definition (HD) maps used by autonomous vehicles is compromised due to inconsistencies in sensor data collected by heterogeneous data-gathering vehicles, which can affect the safety and reliability of autonomous navigation.
Innovation Solution
A machine-learning model is employed to transform sensor data from diverse data-gathering vehicles into a common data space, minimizing discrepancies and enhancing the accuracy of HD maps. This model uses neural networks to encode sensor data into a latent representation and decode it into accurate HD map data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If data are collected by a fleet of vehicles equipped with sensors, then the coverage and data quantity for HD map generation are improved, but the accuracy of the HD map deteriorates due to differences in data-collection equipment
Solution Approach 1:
The patent transforms sensor data from heterogeneous sources by adjusting and normalizing parameters such as spatial coordinates, temporal stamps, and sensor calibration values. This allows data from different vehicle types and sensor configurations to be aligned to a common reference frame, resolving the accuracy deterioration caused by equipment differences while preserving the benefits of large data quantities from fleet-wide collection.
Solution Approach 2:
The system introduces an intermediary processing layer that acts as a mediator between raw sensor data from diverse vehicles and the final HD map generation process. This intermediary layer performs data fusion, coordinate transformation, and quality weighting to harmonize heterogeneous data sources, enabling accurate map generation from fleet-wide data without requiring uniform sensor equipment across all vehicles.
2Ease of operation
If traditional maps are used for navigation, then the simplicity and ease of use are improved, but the safety and reliability of autonomous driving deteriorates as the task is left to human drivers without assistance
Solution Approach 1:
The navigation system is segmented into multiple functional layers: traditional map provision for basic navigation guidance, and HD map-based autonomous driving assistance for safety-critical functions. This segmentation allows the system to maintain ease of use through familiar map interfaces while simultaneously providing enhanced safety through detailed HD maps that enable autonomous vehicles to perceive and respond to road hazards, traffic signs, and environmental conditions that traditional maps cannot represent.
Solution Approach 2:
The system merges traditional two-dimensional maps with three-dimensional HD maps to create a hybrid navigation system. The traditional map layer provides familiar navigation routes and waypoints for user guidance, while the HD map layer overlays detailed spatial information about road geometry, curbs, dividers, and hazards. This combination preserves the ease of use of traditional navigation while adding the safety benefits of comprehensive environmental awareness for autonomous driving operations.
Data Source
AI summary
In one embodiment, a method includes a computing system collecting first sensor data, of a particular geographic location, from a first type of sensor. The computing system may process the first sensor data to identify one or more first objects at the particular geographic location. The computing system may access an existing high-definition (HD) map associated with the particular geographic location. The existing HD map includes one or more second objects and is generating using second sensor data collected with a second type of sensor. The computing system may determine whether the one or more first objects are included in the existing HD map. In response to determining that the one or more first objects are not included in the existing HD map, the computing system may update the existing HD map to generate an updated HD map that includes the first objects and the second objects.