Mobile Unit Pose Estimation Using Collision-Free Region Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing localization methods for mobile units, such as autonomous vehicles, are inaccurate in areas with few or no detectable landmarks, or where unambiguous landmark assignment is not possible.
Innovation Solution
A method that uses a sensor device to capture images of the surroundings and calculates a collision-free area based on these images, comparing it with a drivable area marked on a semantic map to determine the pose of the mobile unit.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If landmark-based localization methods are used, then positioning can be achieved in environments with visible landmarks, but the method fails or becomes inaccurate in areas with few or no detectable landmarks
Solution Approach 1:
The patent introduces an intermediary approach by using semantic map data (drivable area information) as a mediator between the vehicle's sensor observations and its position estimation. Instead of directly relying on landmark detection, the system compares the drivable area derived from sensor data with the drivable area from semantic maps to indirectly determine position, enabling reliable localization in environments with few or no detectable landmarks
2Measurement precision
If landmark detection and assignment methods are used, then position determination is possible when landmarks are visible and assignable, but the method cannot be implemented or has insufficient accuracy when few or no landmarks can be detected or unambiguously assigned
Solution Approach 1:
The patent replaces the mechanical/visual process of landmark detection and assignment with a data-processing approach using semantic maps. Instead of detecting physical landmarks and assigning them to map features, the system processes sensor data to generate drivable area representations and compares these with pre-stored semantic map data, eliminating the need for complex landmark detection and assignment operations
Solution Approach 2:
The patent changes the parameters used for position determination from landmark-based parameters (landmark positions, orientations) to drivable area-based parameters (spatial extent, shape, position of drivable regions). This parameter transformation allows the system to operate in environments where traditional landmark parameters cannot be reliably obtained
Data Source
Figure 1~2
AI summary
The invention relates to a method for determining the pose of a mobile unit (1) with at least one sensor device that is designed to record images of the surroundings (PIC) of the mobile unit (1), and a map (MAP) in which at least one navigable region (2) is marked, wherein at least one image of the surroundings (PIC) is recorded by way of the at least one sensor device and at least one collision-free region (3, 3.1, 3.2) is calculated on the basis of the at least one recorded image of the surroundings (PIC), and the pose of the mobile unit (1) is determined by comparing the at least one calculated collision-free region (3, 3.1, 3.2) with the at least one navigable region (2) marked in the map (MAP). A system for performing the method is furthermore also described.