Resilient State Estimation Using Partial UIOs Under Sensor Attack
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing control systems in Cyber Physical Systems (CPS) are vulnerable to sensor attacks and disturbances, which can lead to inaccurate system state estimation and operational failures.
Innovation Solution
A system state estimation apparatus and method using partial state Unknown Input Observers (UIOs) to generate and combine partial state estimation values, removing the influence of sensor attacks and failures, and determining a final state estimation value through a finite space optimization process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional state estimation methods are used in CPS, then the system can operate with simple architecture, but the system becomes vulnerable to sensor attacks and disturbances leading to inaccurate state estimation
Solution Approach 1:
The system divides the state estimation problem into multiple independent partial state estimation sub-problems, each handled by a separate partial state UIO. Each observer estimates a specific subset of system states independently, and the results are combined to form the complete state estimation. This segmentation allows the system to achieve high reliability under sensor attacks while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The patent introduces an intermediary combination mechanism that integrates multiple partial state estimation values to produce the final state estimation. This intermediary layer filters out corrupted information from attacked sensors by combining estimates from multiple independent observers, thereby improving reliability while keeping individual observer complexity low.
2Measurement precision
If multiple sensors are used to improve observation accuracy, then the system can achieve better state estimation, but the system becomes more vulnerable to sensor attacks and failures
Solution Approach 1:
The system converts the potential harm of having multiple sensors vulnerable to attacks into a benefit by using diverse partial state UIOs that can cross-validate each other. When some sensors are attacked, the unattacked sensors' corresponding partial state estimators provide accurate information that compensates for the corrupted data, thereby converting the vulnerability into a robust estimation capability.
Solution Approach 2:
Each partial state UIO is designed with specialized local quality to estimate specific state variables using specific sensor inputs. This localized design means that an attack on one sensor only affects its corresponding partial estimator, while other partial estimators remain unaffected and can still provide accurate local state information.
3Reliability
If robust state estimation methods are implemented to resist disturbances, then the system achieves higher reliability, but the computation time and complexity increase
Solution Approach 1:
By segmenting the robust estimation task into multiple parallel partial state UIOs, each processing a subset of state variables independently, the system achieves disturbance resistance without requiring sequential complex computations. The parallel architecture reduces overall computation time while maintaining robustness through the combination of partial estimates.
4Reliability
If partial state UIOs are used to remove sensor attack influence, then the system achieves accurate state estimation under attack, but the system architecture becomes more complex
Solution Approach 1:
The patent applies segmentation by creating multiple independent partial state UIOs, each with a specific function to estimate particular state variables. This modular segmentation makes the complex attack-resilient system easier to design, implement, and maintain compared to a single monolithic robust observer, as each module can be developed and validated independently.
Data Source
AI summary
Disclosed are an apparatus and a method for estimating a system state. Accordingly, it is possible to perform Resilient State Estimation (RSE) that is robust to disturbance and is autonomously restored from sensor attack/failure.