Sensor Fusion in Cooperative Perception Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current cooperative perception systems face challenges in improving the accuracy of object classification and detection while efficiently integrating outputs from sensors of different types.
Innovation Solution
The system employs a cooperative perception system that includes multiple imaging sensors connected to machine learning (ML) systems. These ML systems process output images to yield hypotheses with regressed parameters and variation data, which are then fused using variational information to improve hypothesis accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If sensor outputs from multiple sensors are integrated to improve object detection accuracy, then the accuracy of object classification and detection is improved, but the volume of information to be exchanged and processed increases
Solution Approach 1:
The patent extracts only the essential elements from sensor outputs - specifically hypotheses about objects and their associated uncertainty information - rather than processing complete sensor data. This extraction approach allows multiple sensor outputs to be integrated for improved accuracy while keeping the processed information volume manageable by focusing only on detection hypotheses and uncertainty metrics.
Solution Approach 2:
The patent segments the sensor integration process into distinct components: individual sensor hypothesis generation, uncertainty calculation for each sensor, and subsequent fusion of these segmented hypothesis sets. This segmentation allows each sensor to be processed independently first, then combined systematically, improving accuracy through integration while managing information volume through structured processing.
2Measurement precision
If outputs from multiple sensors are integrated to improve detection accuracy, then the accuracy of object classification is improved, but the complexity of the system increases
Solution Approach 1:
The patent introduces uncertainty information as an intermediary element that mediates the integration of multiple sensor outputs. Rather than directly combining complex sensor data, the system uses uncertainty metrics as an intermediate representation that simplifies the fusion process. This intermediary approach improves classification accuracy through multi-sensor integration while reducing system complexity by providing a standardized method for combining hypotheses.
Solution Approach 2:
The patent changes the parameter representation from raw sensor outputs to structured hypotheses with associated uncertainty parameters. By transforming sensor data into this standardized parameter format (hypotheses + uncertainty), the system can integrate multiple sensors more efficiently. This parameter transformation improves classification accuracy while managing complexity through consistent data representation.
3Measurement precision
If complete sensor data is processed to yield accurate hypotheses, then the accuracy of detection is improved, but the processing time and computational resources increase
Solution Approach 1:
The patent applies preliminary action by having each individual sensor generate hypotheses and calculate uncertainty information before the integration step. This preliminary processing allows each sensor to work independently and efficiently on its own data, producing ready-to-fuse results. This approach improves detection accuracy through comprehensive sensor processing while reducing overall processing time by avoiding redundant computations during the integration phase.
Data Source
AI summary
Cooperative perception systems comprise a plurality of imaging sensors that are connected to provide output images to one of one or more machine learning (ML) systems, each ML system is trained to process the output images to yield variational hypotheses. Each of the variational hypotheses comprises one or more objects and, for each of the objects, values for each of a plurality of regressed parameters and variation data indicating uncertainty of the value for each of the plurality of regressed parameters. A processor receives and fuses the hypotheses using the variation data to yield a refined hypothesis. The refined hypothesis may provide an input to a control system for a vehicle, robot or other apparatus.