Feature-Based Localization Maps with On-Demand Geometric Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle localization methods face challenges in rural areas or highways due to insufficient distinguishable semantic features, requiring large memory for geometric features and limited wireless communication transfer.
Innovation Solution
A method that combines semantic and geometric features from multiple sensors to create a digital map, using a control unit to check and update map data, ensuring a robust and precise localization by extracting and complementing features as needed.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If geometric features are extracted to ensure sufficient localization features, then localization reliability is improved, but memory space requirement increases
Solution Approach 1:
The patent segments features into two distinct types: semantic features (first features) and geometric features (second features). This segmentation allows the system to store only semantic features in memory, which require minimal space, while extracting geometric features temporarily from sensor data when needed for localization, thus resolving the contradiction between localization reliability and memory space requirements
Solution Approach 2:
The patent extracts only the necessary semantic features from sensor data and stores them in memory, while geometric features are extracted on-demand from incoming measured data during localization operations. This extraction approach ensures that minimal memory space is occupied while sufficient localization features are available when needed
2Quantity of substance
If semantic features are used to reduce memory space, then memory efficiency is improved, but localization precision deteriorates in rural areas
Solution Approach 1:
The patent merges semantic features (stored in memory) and geometric features (extracted from sensor data) into a hybrid feature-based localization system. This combination allows the system to benefit from both the memory efficiency of semantic features and the precision of geometric features, particularly in rural areas where semantic features alone may be insufficient
Solution Approach 2:
The patent implements a dynamic feature extraction and selection mechanism that adapts to different environments. In urban areas with rich semantic features, the system relies primarily on stored semantic data, while in rural areas with fewer semantic features, it dynamically extracts and utilizes geometric features from current sensor measurements to maintain localization precision
3Measurement precision
If geometric features are extracted to improve localization precision, then localization precision is improved, but wireless communication transfer capability is exceeded
Solution Approach 1:
The patent extracts only the essential semantic features from sensor data and stores them in memory, while geometric features are extracted temporarily from incoming measured data during localization operations. This extraction approach ensures that minimal data is transferred via wireless communication, preventing communication link overload while maintaining sufficient localization precision
Data Source
AI summary
A method is provided for creating at least one map of vehicle surroundings with the aid of a control unit. It is checked based on a comparison between received measured data and stored or received map data, whether first features, for example semantic features, are present and complete. First features available in a vehicle surroundings are extracted from the received measured data if no or incomplete map data are present. It is checked whether a localization is possible within the vehicle surroundings with the aid of the first semantic features. If a localization is imprecise or not possible with the aid of the ascertained first features, second features are extracted from the received measured data. A digital map of vehicle surroundings is created based on the ascertained first features and/or the second features. Furthermore, a control unit, a computer program as well as a machine-readable memory medium are provided.