Radar Localization Map Creation Using Satellite Transfer Model
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for creating vehicle localization maps require extensive travel by mapping vehicles, which is inefficient and costly, and do not efficiently utilize high-frequency satellite data for keeping maps up-to-date.
Innovation Solution
A method utilizing a transfer model that converts photographic satellite data into surround-field sensor data, allowing for the creation of localization maps using satellite data alone, reducing the need for extensive vehicle travel and enabling rapid updates.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If mapping vehicles travel extensively through areas to be mapped using conventional methods, then localization maps can be created with sufficient accuracy, but the process becomes inefficient and costly in terms of time and resources
Solution Approach 1:
The patent uses satellite-based photographic data as a copy or representation of the real-world terrain, replacing the need for direct physical surveying by mapping vehicles. The satellite images serve as a surrogate that can be processed to create localization maps without requiring extensive ground truth data collection through vehicle travel.
Solution Approach 2:
The patent replaces the mechanical process of physical surveying and ground-based mapping with a remote sensing approach using satellite photography and automated image processing algorithms. This substitution eliminates the need for mapping vehicles to physically traverse and measure every area, dramatically improving efficiency while maintaining accuracy through computational methods.
2Reliability
If conventional mapping methods are used, then localization maps can be created, but they do not efficiently utilize high-frequency satellite data for keeping maps up-to-date
Solution Approach 1:
The patent performs preliminary processing of satellite photographic data to extract and pre-process relevant features, landmarks, and spatial information before integrating them into the localization map system. This preliminary action enables the system to efficiently incorporate high-frequency satellite updates without requiring complete remapping, thus maintaining map currency while utilizing satellite data effectively.
Solution Approach 2:
The patent implements a feedback mechanism where newly acquired satellite data is continuously compared with existing localization maps, and updates are automatically integrated. This feedback loop ensures that the maps remain up-to-date by incorporating the latest satellite imagery information, efficiently utilizing high-frequency satellite data for ongoing map maintenance and updates.
3Area of stationary object
If extensive vehicle travel is required for map creation, then comprehensive coverage can be achieved, but the logistical and technical effort increases significantly
Solution Approach 1:
The patent transitions from two-dimensional ground-based mapping to three-dimensional spatial understanding by incorporating satellite imagery that captures elevation, terrain features, and aerial perspectives. This dimensional change enables comprehensive coverage of large areas without requiring physical vehicle travel, as the satellite data provides a top-down view that can be processed to create accurate localization maps of the entire region.
Solution Approach 2:
The patent uses satellite photographic data that serves multiple functions simultaneously: it provides overview coverage of large areas, captures detailed features for localization, and can be processed to generate various map types. This multi-functionality eliminates the need for specialized mapping vehicles to perform multiple tasks, reducing logistical and technical complexity while achieving comprehensive coverage.
Data Source
AI summary
A method for creating a radar localization map. The method includes: a) sensing a defined region using a surround-field sensor of a mapping vehicle; b) providing photographic satellite data of the defined region with the aid of a satellite, c) determining matching detected objects of the region in the surround-field sensor data and in the photographic satellite data; d) generating a transfer model from the matching detected objects, the photographic satellite data being transferable reciprocally into the surround-field sensor data utilizing the transfer model; and e) creating the radar localization map by use of the transfer model using photographic satellite data, the photographic satellite data being converted into corresponding data of the radar localization map.