Autonomous Driving Map Patch Selection for Accurate Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle navigation systems face challenges in efficiently downloading and utilizing map data for autonomous driving, particularly in limited communication bandwidth and accuracy of positioning, especially when landmarks are difficult to detect due to environmental factors.
Innovation Solution
A system that distributes map data in patch units from a server to a vehicle, using a processor to acquire the vehicle's position via positioning sensors and download relevant map patches based on the vehicle's current location, incorporating image processing from cameras to detect landmarks and update map data, and employing sensor fusion for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If map data is downloaded in full from the server, then complete map information is available for autonomous driving, but communication bandwidth requirements increase significantly
Solution Approach 1:
The patent divides map data into multiple patches corresponding to different geographic areas. The server stores and manages map data as a collection of discrete patches, and the vehicle downloads only the specific patches relevant to its current location and planned route, rather than downloading the entire map dataset. This segmentation approach reduces communication bandwidth requirements while ensuring the vehicle has access to complete map information for its operational area.
2Reliability
If the vehicle downloads map patches proactively before reaching an area, then map data is available when needed, but storage requirements and data management complexity increase
Solution Approach 1:
The system proactively downloads map patches to the vehicle's storage device before the vehicle reaches the corresponding geographic areas. The server determines which patches to download based on the vehicle's current position and planned route, and initiates the download process in advance. This preliminary action ensures that map data is available when needed for autonomous driving operations, while the server manages the complexity of data selection and delivery.
3Measurement precision
If positioning accuracy is improved using multiple sensors and landmark detection, then vehicle location is determined more precisely, but system complexity and computational requirements increase
Solution Approach 1:
The patent combines multiple positioning approaches by merging sensor fusion (using accelerometer, gyroscope, and other vehicle sensors) with landmark detection (using camera images to identify and locate landmarks). The server receives both sensor-based position information and landmark-based position information, integrates these data sources, and determines the vehicle's position by combining the results. This merging approach improves positioning accuracy while distributing computational complexity across the vehicle and server.
Solution Approach 2:
The server acts as an intermediary that receives position information from multiple sources (sensor fusion results and landmark detection results), processes and integrates this information, and determines the vehicle's position. Rather than requiring the vehicle to perform complex integration of multiple positioning methods, the server mediates the fusion process, reducing the computational burden on the vehicle while achieving high positioning accuracy.
Data Source
AI summary
A system for autonomously driving a vehicle along a road segment using map data including coordinate information of a landmark disposed along a road, includes at least one processor. The map data is distributable from a predetermined server to the vehicle in patch units, which are data sets for different areas. The processor is configured to: acquire a position of the vehicle based on a detection result of a positioning sensor mounted on the vehicle; and set a patch as a download target including a road which has a possibility that the vehicle passes through and is determined based on a current position of the vehicle.