Autonomous Vehicle Localization Input Filtering Using Occupancy Maps
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Autonomous vehicles face challenges in accurately localizing their position due to high computational demands and the risk of incorrect measurements, especially when sensors detect the driving surface as an obstacle, leading to increased processing requirements and potential collisions.
Innovation Solution
A filtering procedure that uses a predefined occupancy card and an optimized data structure, such as a K-D tree, to discard incorrect measurement values by assigning them to a search radius and checking for occupied cells within that radius, thereby reducing data processing and enhancing localization robustness.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a high frequency of localization processes and acquisition of a large number of sensor data are used to achieve high-precision localization, then localization precision is improved, but computational requirements increase
Solution Approach 1:
The patent applies preliminary action by pre-processing sensor data through filtering before localization. The filtering step removes erroneous measurements in advance, so that only valid data points are processed during the computationally intensive localization algorithm execution. This preliminary filtering reduces the volume of data requiring full localization processing while maintaining localization precision.
Solution Approach 2:
The patent extracts and removes erroneous measurements from the sensor data stream before feeding data to the localization algorithm. By identifying and discarding invalid data points (such as those from sensor tilt or ground reflections), the system reduces computational load on the localization processor while preserving the quality of remaining data for accurate localization.
2Quantity of substance
If sensor units detect all objects in the environment including the driving surface, then measurement completeness is improved, but incorrect measurements increase
Solution Approach 1:
The patent applies local quality by applying different validation criteria to different types of measurements. Rather than treating all detected objects uniformly, the system uses spatial reasoning and occupancy map comparison to identify which measurements are likely valid (objects at appropriate distances and angles) and which are erroneous (ground surface detections when sensors are tilted). This selective validation improves measurement reliability while preserving complete object detection.
Solution Approach 2:
The patent converts the harmful effect of erroneous measurements into a benefit by using the occupancy map and spatial context to identify invalid detections. The presence of additional measurements, including erroneous ones, provides more data points for the filtering algorithm to work with, and the systematic validation process transforms this excess data into reliable information by eliminating false positives.
3Quantity of substance
If erroneous measured values are not filtered out, then data completeness is improved, but localization reliability deteriorates
Solution Approach 1:
The patent performs preliminary validation of measured values against the occupancy map and spatial constraints before feeding data to the localization algorithm. This pre-filtering step maintains data completeness by preserving all potentially valid measurements while removing obvious errors, thereby improving localization reliability without losing useful information.
Solution Approach 2:
The patent implements feedback by comparing sensor measurements against the pre-built occupancy map of the environment. Measurements that contradict the known environment (such as detecting ground surface in directions where the map shows open space) are flagged as erroneous and discarded. This feedback mechanism maintains reliability while preserving complete valid data for localization.
Data Source
Figure 1
Figure 2~3
AI summary
The present invention relates to a method for filtering inputs to a localization method of an autonomous vehicle (10) in an operating environment (U), comprising the steps of providing a predefined occupancy map representing known objects (W) present in the operating environment (U), generating an optimized data structure representing occupied cells in a representation of the occupancy map in a coordinate system, acquiring the operating environment (U) of the vehicle (10) by means of at least one sensor unit (12) configured to detect objects in the operating environment (U) and their distances from the vehicle (10), wherein each detected object is assigned a measured value (M1, M2) in the coordinate system, and searching within a search radius around the measured value (M1, M2) for an occupied cell in the optimized data structure.If an occupied cell is found within the search radius: the measured value (M1) is passed on to a subsequent localization procedure; and if no occupied cell is found within the search radius, the measured value (M2) is discarded.