scholarly journals Synthesis-guided Adversarial Scenario Generation for Gray-box Feedback Control Systems with Sensing Imperfections

2021 ◽  
Vol 20 (5s) ◽  
pp. 1-25
Author(s):  
Liren Yang ◽  
Necmiye Ozay

In this paper, we study feedback dynamical systems with memoryless controllers under imperfect information. We develop an algorithm that searches for “adversarial scenarios”, which can be thought of as the strategy for the adversary representing the noise and disturbances, that lead to safety violations. The main challenge is to analyze the closed-loop system's vulnerabilities with a potentially complex or even unknown controller in the loop. As opposed to commonly adopted approaches that treat the system under test as a black-box, we propose a synthesis-guided approach, which leverages the knowledge of a plant model at hand. This hence leads to a way to deal with gray-box systems (i.e., with known plant and unknown controller). Our approach reveals the role of the imperfect information in the violation. Examples show that our approach can find non-trivial scenarios that are difficult to expose by random simulations. This approach is further extended to incorporate model mismatch and to falsify vision-in-the-loop systems against finite-time reach-avoid specifications.

2020 ◽  
Vol 30 (17) ◽  
pp. 7103-7129
Author(s):  
Xiaodong Xu ◽  
Jodie M. Simkoff ◽  
Michael Baldea ◽  
Leo H. Chiang ◽  
Ivan Castillo ◽  
...  

1989 ◽  
Vol 111 (2) ◽  
pp. 339-342
Author(s):  
R. Shoureshi

Closed-loop control systems, especially linear quadratic regulators (LQR), require feedbacks of all states. This requirement may not be feasible for those systems which have limitations due to geometry, power, required sensors, size, and cost. To overcome such requirements a passive method for implementation of state feedback control systems is presented.


1969 ◽  
Vol 91 (2) ◽  
pp. 246-249 ◽  
Author(s):  
W. F. Horton ◽  
C. T. Leondes

Definitions of system sensitivity for linear single variable systems have been extended, in the past, to linear multivariable systems in the form of a sensitivity matrix. The role of the sensitivity matrix in multivariable feedback control systems is studied further in this paper. The sensitivity matrix serves the dual function of governing the effects of plant parameter variation on the system transfer matrix and governing the effects of disturbances on the system output. The design implications of this are considered and it is shown that certain controllability/observability conditions are necessary if the system design is to be effective. By appropriate design of the loop gain matrix, L(s), a desired insensitivity to system error sources may be achieved. Unless the system has certain controllability/observability properties insensitivity cannot be achieved. It is shown that L(s) must have the property of functional reproducibility which is a relatively strong controllability/observability requirement.


Sign in / Sign up

Export Citation Format

Share Document