Physics-Guided Model Conformance for Unknown-Unknown Detection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Safety-critical cyber-physical systems face challenges in detecting 'unknown-unknown' errors that arise from unpredictable human interactions and complex real-world scenarios, as existing design-time safety assurance approaches cannot account for these unforeseen conditions, leading to potential safety violations and accidents.

Innovation Solution

A method that uses Physics Guided Surrogate Modeling (PGSM) to derive and evaluate model coefficients based on operational traces, identifying error time steps by comparing post-deployment coefficients against a conformal range defined by pre-deployment model coefficients, ensuring safety conditions are met through Signal Temporal Logic functions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If design-time safety assurance approaches are used, then safety conditions can be verified during development, but unknown-unknown errors from unpredictable real-world scenarios cannot be detected

Engineering Contradiction:

Improvesafety assuranceVSAvoidhandling of unknown scenarios

Core Design Contradiction:

ReliabilityVSAdaptability or versatility

Solution Approach 1:

The system performs preliminary actions by establishing a conformal range during the design phase based on pre-deployment model coefficients. This pre-established range serves as a baseline for detecting deviations caused by unknown-unknown errors during runtime, enabling the system to adapt to unforeseen scenarios while maintaining safety guarantees.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The system implements continuous feedback by monitoring post-deployment model coefficients and comparing them against the pre-established conformal range. When deviations exceed the conformal range, the system identifies potential unknown-unknown errors and triggers safety responses, creating a closed-loop feedback mechanism that adapts to real-world variations.

Inventive Principle:

Principle #23Feedback

2Reliability

If pre-deployment model coefficients are used for safety verification, then safety conditions can be ensured, but deviations from actual post-deployment behavior may go undetected

Engineering Contradiction:

Improvesafety condition verificationVSAvoiddetection of model deviations

Core Design Contradiction:

ReliabilityVSMeasurement precision

Solution Approach 1:

The system monitors changes in model coefficients as key parameters that characterize system behavior. By tracking how post-deployment coefficients deviate from pre-deployment values and comparing against the conformal range, the system achieves precise detection of behavioral changes while maintaining safety verification through the structured parameter comparison framework.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The system replaces traditional mechanical model verification with a statistical conformance approach. Instead of relying solely on rigid pre-deployment model matching, the system uses probabilistic conformal ranges derived from operational traces to allow for natural variations while still detecting significant deviations indicative of unknown-unknown errors.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Reliability

If conformal range monitoring is implemented, then unknown-unknown errors can be detected early, but system complexity increases

Engineering Contradiction:

Improveerror detection capabilityVSAvoidmonitoring system structure

Core Design Contradiction:

ReliabilityVSDevice complexity

Solution Approach 1:

The system introduces an intermediary conformal range as a mediator between pre-deployment safety guarantees and post-deployment runtime behavior. This conformal range acts as a buffer zone that absorbs normal variations while flagging significant deviations, simplifying the monitoring logic to a straightforward comparison operation rather than complex real-time analysis.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The system creates a simplified copy of the pre-deployment model coefficients and establishes a conformal range around them. This copied reference model with its conformal bounds serves as a lightweight monitoring mechanism that can be efficiently compared against runtime behavior without requiring full complexity of the original safety verification framework.

Inventive Principle:

Principle #26Copying

Data Source

PatentUS20240427321A1Systems and methods for detection of unknown-unknowns in dynamical systems using statistical conformance with physics-guided process models

Publication Date: 2024.12.26 THE ARIZONA BOARD OF REGENTS ON BEHALF OF THE UNIV OF ARIZONA

AI summary

A framework includes a system and associated computer-implemented methods for detecting behavioral changes in a dynamical system that can lead to unsafe conditions before an output of the dynamical system violates a safety threshold, especially for dynamical systems with unmodeled inputs and unmodeled dynamics. In particular, the framework aims to detect “unknown-unknown” errors that may be present in a post-deployment model of the dynamical system that may not be anticipated or modellable by its designers, and are often not directly observable through input-output traces. This is achieved by evaluating conformance of post-deployment model coefficients of the post-deployment model with respect to a set of pre-deployment (ideal) model coefficients. The framework can estimate a future time step where the output of the dynamical system is expected to violate a safety violation based on the post-deployment model.