Logical Representation of Sensor Data for Autonomous Driving
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for abstracting sensor data into logical scenarios for testing automated and autonomous driving systems fail to accurately preserve the overall situation while allowing for variable changes, resulting in either insufficient data segments or excessive parameters, leading to poor accuracy and generalization.
Innovation Solution
A method using a trained machine learning algorithm to transform sensor data into a reduced complexity logical representation, minimizing the number of classes and optimizing parameters to preserve the original scenario's position, time, and speed, while enabling simple variable changes for generating new test cases.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If segmentation algorithms are used to generalize sensor data into logical scenarios, then data abstraction is achieved, but the accuracy of trajectory representation deteriorates when using a small number of data segments
Solution Approach 1:
The patent applies dynamic segmentation by adapting the number and boundaries of data segments based on the complexity of the driving scenario. The segmentation is not fixed but dynamically adjusted to preserve important trajectory characteristics while achieving effective generalization. This resolves the contradiction by making the segmentation strategy flexible rather than static.
Solution Approach 2:
The patent changes the parameters of segmentation (number of segments, segment boundaries, aggregation levels) to optimize both abstraction quality and trajectory accuracy. By adjusting these parameters based on scenario requirements, the system achieves high-level abstraction without sacrificing essential accuracy.
2Measurement precision
If a large number of data segments are used to improve trajectory accuracy, then measurement precision improves, but device complexity and the number of parameters to be changed increases
Solution Approach 1:
The patent applies hierarchical segmentation where sensor data is divided into meaningful segments that capture essential scenario characteristics. By segmenting data at appropriate levels of granularity, the system achieves good trajectory representation without requiring an excessive number of fine-grained segments, thus balancing accuracy with complexity.
Solution Approach 2:
The patent merges adjacent data segments that share similar characteristics into aggregated segments. This merging process reduces the total number of segments while preserving the essential trajectory information, thereby reducing complexity without significantly compromising accuracy.
3Ease of manufacture
If initial settings of variable values are used in segmentation algorithms, then the processing is simplified, but the accuracy of representing the actual situation deteriorates
Solution Approach 1:
The patent incorporates feedback mechanisms where the segmentation algorithm iteratively adjusts variable settings based on how well the segmented data represents the original sensor data. This feedback loop allows the system to start with simple initial settings but refine them to achieve accurate scenario representation.
Solution Approach 2:
The patent performs preliminary analysis of the sensor data to determine appropriate variable settings before executing the main segmentation process. This preliminary action ensures that the segmentation variables are optimized for the specific data characteristics, improving accuracy without adding complex real-time adjustments.
4Adaptability or versatility
If more parameters are changed to vary the scenario for generating new test cases, then adaptability improves, but the complexity of scenario modification increases
Solution Approach 1:
The patent creates logical scenarios with universal structures that can represent multiple driving situations through a limited set of parameter variations. By designing a universal scenario template, the system can generate diverse test cases by modifying only key parameters rather than changing many individual variables, thus improving adaptability while controlling complexity.
Data Source
AI summary
A method and system for generating a reduced complexity logical representation of a data set of sensor data, having a using of an algorithm on the second data set for reducing the complexity of the logical scenario, and an outputting of a third data set representing a reduced complexity logical scenario of the second data set. The invention additionally relates to a method for providing a trained machine learning algorithm for generating a reduced complexity representation of a data set of sensor data.