Cross-Sensor Patch Mapping for Accurate Vehicle Sensor Fusion
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing sensor systems for autonomous vehicles and advanced driving assistance systems face inaccuracies due to misalignment and differing operations, necessitating an efficient and accurate method to combine information from multiple sensors.
Innovation Solution
A method for sensor fusion using a correlation function developed through supervised machine learning, which determines which patches of sensed information units from different sensors should be fused, reducing computational resources by selecting relevant patches for correlation and applying a projective transformation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If information from different types of sensors is combined, then the accuracy and completeness of environmental assessment is improved, but the computational complexity and resource consumption increase
Solution Approach 1:
The patent divides the sensor data processing into discrete patches or regions of interest. Instead of processing entire sensor inputs globally, the system segments the environmental assessment into multiple local patches that can be processed independently and in parallel, reducing computational complexity while maintaining assessment accuracy.
Solution Approach 2:
The patent applies different processing strategies to different patches based on their specific characteristics. High-priority patches requiring precise fusion of multiple sensor types receive more computational resources, while low-priority patches use simplified processing, optimizing the balance between accuracy and computational load.
2Productivity
If patches from different sensors are selected for fusion, then the computational resources are reduced, but the accuracy of sensor fusion may deteriorate
Solution Approach 1:
The patent performs preliminary selection and prioritization of patches before fusion. By pre-identifying which patches contain critical information for autonomous driving decisions, the system can focus computational resources on these high-value patches, ensuring accuracy is maintained where it matters most while improving overall efficiency.
Solution Approach 2:
The patent incorporates feedback mechanisms that evaluate the quality and relevance of selected patches. This feedback loop allows the system to adjust patch selection criteria based on environmental conditions, sensor performance, and driving context, ensuring that the most informative patches are fused while maintaining computational efficiency.
Data Source
AI summary
A computer-implemented method for sensor fusion in relation to at least partially autonomous driving of a vehicle. The method may include obtaining first signatures of first patches of a first type sensed information unit (SIU) that was sensed by a first sensor of a first type; obtaining second signatures of second patches of a second type SIU that was sensed by a second sensor of a second type, the second type differs from the first type; wherein the first sensor and the second sensor are associated with the vehicle; finding correlations by applying a correlation function between the first signatures and the second signatures; wherein the finding is executed by a mapping system; and determining, based on the correlations and by the mapping system, a mapping between the first patches and the second patches, the mapping to be used in an at least partially autonomous driving of a vehicle; wherein the correlation function having been developed by applying a supervised machine learning process based on relationships between members of training signature pairs, each training signature pair comprises a first sensor training signature and a second sensor training signature of a same object.