Ultrasonic Sensor Sensing With OOD Conformity Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing vehicle environment sensing systems face challenges in securing machine learning models against out-of-distribution (OOD) data and adversarial attacks, which can compromise vehicle safety by misleading algorithms.
Innovation Solution
A method and device that utilize a trained machine learning model to calculate uncertainty measures and conformity scores, incorporating deviation values to generate prediction sets, ensuring secure vehicle reactions by generating control signals based on these scores, and initiating emergency maneuvers if necessary.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If machine learning models are used for environment sensing in vehicles, then sensing accuracy and automation are improved, but the system becomes vulnerable to out-of-distribution data and adversarial attacks
Solution Approach 1:
The patent applies preliminary action by calculating conformity scores and uncertainty measures before making final sensing decisions. The system pre-processes sensor data through conformity functions that compare incoming data against training data distributions, identifying potential adversarial examples before they can compromise the machine learning model's reliability
Solution Approach 2:
The patent introduces an intermediary conformity assessment layer between the sensor input and the machine learning model. This intermediary calculates conformity scores that mediate between raw sensor data and model processing, filtering out out-of-distribution data points that could otherwise misleading the automation system
2Reliability
If conformity functions and uncertainty measures are calculated to secure against OOD data, then reliability is improved, but device complexity increases
Solution Approach 1:
The patent applies universality by designing the conformity function to serve multiple purposes simultaneously: it assesses data conformity, calculates uncertainty measures, and identifies out-of-distribution points all through a single integrated function, reducing the need for separate complex verification systems
Solution Approach 2:
The patent uses parameter changes by transforming the conformity function's output parameters to represent different aspects of data reliability. By adjusting how conformity scores are calculated and interpreted, the system adapts to different sensing scenarios without requiring fundamentally different system architectures
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
Enhances the reliability of vehicle environment sensing by accurately identifying and responding to potentially manipulated or OOD data points, ensuring safe vehicle operation through calibrated confidence and risk management.
Implementation Method 1
recording environment data by means of at least one vehicle sensor
Implementation Method 2
ultrasonic sensors
Data Source
AI summary
The present disclosure relates to a method, to a computer program comprising instructions, and to a device for environment sensing in a vehicle. For the environment sensing, environment data are recorded by means of at least one vehicle sensor. A prediction is calculated by means of a trained machine learning model based on the recorded environment data, wherein the prediction includes a measure of uncertainty for the prediction. Equally, a deviation value is determined for the recorded environment data that provides a measure of how significantly the recorded environment data deviate from training data for the machine learning model. At least one conformity score is determined based on the measure of uncertainty and the deviation value and then a prediction set is determined based on the at least one determined conformity score. Then, a control signal is generated depending on the determined prediction set.