Vehicle Environment Map Fusion Using Object Orientation Cues
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle surroundings detection systems fail to accurately account for measurement artifacts, leading to incorrect deletion of objects from the environment map due to unfavorable surface orientations, which results in misinterpretation of free areas and reduced object existence probabilities.
Innovation Solution
Assigning orientation attributes to recognized objects, such as surface orientations, during data fusion, and using these attributes to determine if the probability of existence should be reduced based on the difference angle between the measurement angle and the surface normal, thereby discarding measurement artifacts that indicate free areas due to unfavorable reflections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If objects are deleted from the environment map when free areas are measured at their position, then the environment map is updated to reflect current sensor data, but measurement artifacts cause incorrect deletion of objects that still exist due to unfavorable surface orientations
Solution Approach 1:
The system performs preliminary actions by storing historical measurement data and object information in the environment map before new measurements are processed. When a free area is measured, the system first checks the environment map for previously stored objects at that position, and only deletes the object if the measurement is confirmed to be valid (not an artifact). This preliminary checking mechanism prevents incorrect deletion due to measurement artifacts.
Solution Approach 2:
The system uses feedback by continuously comparing new sensor measurements with stored environment map data. When a discrepancy is detected (free area measured where an object should exist), the system evaluates whether this represents a real change or a measurement artifact by referencing historical data and object properties. This feedback loop maintains object existence probability despite temporary measurement errors.
2Adaptability or versatility
If the probability of object existence is reduced when free areas are measured, then the system adapts to changing environments, but objects with unfavorable surface orientations are incorrectly marked as non-existent
Solution Approach 1:
The system segments the detection space into multiple measurement areas and processes each area independently with its own validity evaluation. By dividing the environment into discrete cells or regions, the system can apply different validity assessments to different areas, allowing adaptive response to real environmental changes while maintaining objects that show signs of measurement artifacts in specific segments.
Solution Approach 2:
The system dynamically adjusts object existence probability based on multiple factors including measurement history, object properties, and environmental context. Rather than applying a static threshold, the system uses dynamic evaluation that considers the temporal pattern of measurements and characteristics of the detected object, allowing flexible adaptation to environmental changes while filtering out measurement artifacts.
3Speed
If measurement data is used to update the environment map without validation, then real-time detection is achieved, but measurement artifacts lead to misinterpretation of free areas
Solution Approach 1:
The system applies partial validation by checking only critical aspects of measurement data before updating the environment map. Rather than performing exhaustive validation of all measurement parameters, the system focuses on key indicators of measurement validity (such as consistency with historical data and physical plausibility), enabling fast detection while filtering out obvious measurement artifacts through targeted validation checks.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
This approach improves the accuracy of the environment map by preventing misinterpretation of free areas and maintaining the correct probability of object existence, enhancing the reliability and safety of vehicle assistance systems.
Implementation Method 1
when the surface of an object is aligned in such a way that a large part of the power emitted by a sensor is reflected away from the sensor rather than back towards the sensor
Data Source
Figure 1
Figure 2
Figure 3
AI summary
The invention relates to a method for improved data fusion during environment detection in a motor vehicle (50), comprising the following steps: detecting (101) an environment (40) of the motor vehicle (50) using at least one sensor (2); recognising (102) objects (81) in measurement data (20) detected by the at least one sensor (2); fusioning (115) of the recognised objects (81) and the object positions thereof in an environment map (80), wherein a probability of existence is/will be assigned to objects (81) recorded in the environment map (80), and wherein the probability of existence of a recorded object (81) is reduced, when a free region (86, 86.1) is measured during the detection of the position of the recorded object, wherein a respective orientation attribute (60) is assigned to the objects (81) recorded in the environment map (80) and recorded in the environment map (80), and for a region in the sensor region (83.1, 83.2, 83.3) of the at least one sensor (2), in which a free region (86, 86.1) is measured, the probability of existence of an object (81) recorded in the region of the environment map (80) corresponding to the region is reduced in accordance with the assigned orientation attribute (60). The invention also relates to the associated device (1).