Weiming Xiang

Weiming Xiang
Augusta University · Schoo of Computer and Cyber Sciences

Ph.d

About

94
Publications
14,535
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
2,440
Citations
Introduction
Dr. Xiang is currently an Assistant Professor in the School of Computer and Cyber Sciences at Augusta University. Dr. Xiang's research interest is developing formal synthesis and verification techniques and software tools for cyber-physical systems (CPS). His current research centers on formal methods on safety, security and reliablity of learning-enabled CPS.
Additional affiliations
September 2019 - present
Augusta University
Position
  • Professor (Assistant)
August 2019 - present
Augusta University
Position
  • Professor (Assistant)
September 2016 - July 2019
Vanderbilt University
Position
  • PostDoc Position
Education
September 2009 - June 2014
Southwest Jiaotong University
Field of study
  • Transportation Planning and Management
September 2005 - July 2007
Nanjing University of Science and Technology
Field of study
  • System Engineering
September 2001 - July 2005
East China JiaoTong University
Field of study
  • Automation

Publications

Publications (94)
Article
Verification has emerged as a means to provide formal guarantees on learning-based systems incorporating neural network before using them in safety-critical applications. This paper proposes a new verification approach for deep neural networks (DNNs) with piecewise linear activation functions using reachability analysis. The core of our approach is...
Article
Full-text available
This paper presents an overview survey of verification techniques for autonomous systems, with a focus on safety-critical autonomous cyber-physical systems (CPS) and subcomponents thereof. Autonomy in CPS is enabled by recent advances in artificial intelligence (AI) and machine learning (ML) through approaches such as deep neural networks (DNNs), e...
Chapter
Full-text available
This paper presents the Neural Network Verification (NNV) software tool, a set-based verification framework for deep neural networks (DNNs) and learning-enabled cyber-physical systems (CPS). The crux of NNV is a collection of reachability algorithms that make use of a variety of set representations, such as polyhedra, star sets, zonotopes, and abst...
Chapter
Full-text available
Convolutional Neural Networks (CNN) have redefined state-of-the-art in many real-world applications, such as facial recognition, image classification, human pose estimation, and semantic segmentation. Despite their success, CNNs are vulnerable to adversarial attacks, where slight changes to their inputs may lead to sharp changes in their output in...
Article
The vulnerability of artificial intelligence (AI) and machine learning (ML) against adversarial disturbances and attacks significantly restricts their applicability in safety-critical systems including cyber-physical systems (CPS) equipped with neural network components at various stages of sensing and control. This article addresses the reachable...
Preprint
Full-text available
The vulnerability of artificial intelligence (AI) and machine learning (ML) against adversarial disturbances and attacks significantly restricts their applicability in safety-critical systems including cyber-physical systems (CPS) equipped with neural network components at various stages of sensing and control. This paper addresses the reachable se...
Preprint
Full-text available
Convolutional Neural Networks (CNN) have redefined the state-of-the-art in many real-world applications, such as facial recognition, image classification, human pose estimation, and semantic segmentation. Despite their success, CNNs are vulnerable to adversarial attacks, where slight changes to their inputs may lead to sharp changes in their output...
Preprint
Full-text available
This paper presents the Neural Network Verification (NNV) software tool, a set-based verification framework for deep neural networks (DNNs) and learning-enabled cyber-physical systems (CPS). The crux of NNV is a collection of reachability algorithms that make use of a variety of set representations, such as polyhedra, star sets, zonotopes, and abst...
Conference Paper
Full-text available
This paper presents the Neural Network Verification (NNV) software tool, a set-based verification framework for deep neural networks (DNNs) and learning-enabled cyber-physical systems (CPS). The crux of NNV is a collection of reachability algorithms that make use of a variety of set representations, such as polyhedra, star sets, zonotopes, and abst...
Conference Paper
Full-text available
Convolutional Neural Networks (CNN) have redefined the state-of-the-art in many real-world applications, such as facial recognition, image classification, human pose estimation, and semantic segmentation. Despite their success, CNNs are vulnerable to adversarial attacks, where slight changes to their inputs may lead to sharp changes in their output...
Preprint
Full-text available
Deep neural networks have been widely applied as an effective approach to handle complex and practical problems. However, one of the most fundamental open problems is the lack of formal methods to analyze the safety of their behaviors. To address this challenge, we propose a parallelizable technique to compute exact reachable sets of a neural netwo...
Chapter
This paper proposes novel reachability algorithms for both exact (sound and complete) and over-approximation (sound) analysis of deep neural networks (DNNs). The approach uses star sets as a symbolic representation of sets of states, which are known in short as stars and provide an effective representation of high-dimensional polytopes. Our star-ba...
Preprint
Full-text available
Safety-critical distributed cyber-physical systems (CPSs) have been found in a wide range of applications. Notably, they have displayed a great deal of utility in intelligent transportation, where autonomous vehicles communicate and cooperate with each other via a high-speed communication network. Such systems require an ability to identify maneuve...
Chapter
Full-text available
Reachability analysis is a fundamental problem for safety verification and falsification of Cyber-Physical Systems (CPS) whose dynamics follow physical laws usually represented as differential equations. In the last two decades, numerous reachability analysis methods and tools have been proposed for a common class of dynamics in CPS known as ordina...
Conference Paper
This paper proposes novel reachability algorithms for both exact (sound and complete) and over-approximation (sound) analysis for deep neural networks (DNNs). The approach uses star sets as a symbolic representation of sets of states, which are known in short as stars and provide an effective representation of high-dimensional polytopes. Our star-b...
Conference Paper
Full-text available
This paper proposes novel reachability algorithms for both exact (sound and complete) and over-approximation (sound) analysis for deep neural networks (DNNs). The approach uses star sets as a symbolic representation of sets of states, which are known in short as stars and provide an effective representation of high-dimensional polytopes. Our star-b...
Chapter
Safety-critical distributed cyber-physical systems (CPSs) have been found in a wide range of applications. Notably, they have displayed a great deal of utility in intelligent transportation, where autonomous vehicles communicate and cooperate with each other via a high-speed communication network. Such systems require an ability to identify maneuve...
Conference Paper
Full-text available
Safety-critical distributed cyber-physical systems (CPSs) have been found in a wide range of applications. Notably, they have displayed a great deal of utility in intelligent transportation, where autonomous vehicles communicate and cooperate with each other via a high-speed communication network. Such systems require an ability to identify maneuve...
Conference Paper
Full-text available
Artificial neural networks (ANN) have displayed considerable utility in a wide range of applications such as image processing, character and pattern recognition, self-driving cars, evolutionary robotics, and non-linear system identification and control. While ANNs are able to carry out complicated tasks efficiently, they are susceptible to unpredic...
Chapter
Neural networks have been widely used to solve complex real-world problems. Due to the complex, nonlinear, non-convex nature of neural networks, formal safety and robustness guarantees for the behaviors of neural network systems are crucial for their applications in safety-critical systems. In this paper, the reachable set estimation and safety ver...
Preprint
Full-text available
This paper presents a specification-guided safety verification method for feedforward neural networks with general activation functions. As such feedforward networks are memoryless, they can be abstractly represented as mathematical functions, and the reachability analysis of the neural network amounts to interval analysis problems. In the framewor...
Article
Full-text available
In this paper, a novel conception called virtual clock, which is defined by an artificial timer over a finite cycle, is introduced for stability analysis of discrete-time switched linear systems under minimum dwell-time constraint. Two necessary and sufficient conditions associated with a virtual clock with a sufficient length are proposed to ensur...
Preprint
Full-text available
This survey presents an overview of verification techniques for autonomous systems, with a focus on safety-critical autonomous cyber-physical systems (CPS) and subcomponents thereof. Autonomy in CPS is enabling by recent advances in artificial intelligence (AI) and machine learning (ML) through approaches such as deep neural networks (DNNs), embedd...
Article
In this paper, the stability analysis and H∞ control problems for a class of continuous-time switched systems with a random switching signal are addressed. The dwell time of the switched systems is characterized with a fixed part and a random part. A new stability criterion, which is equivalent to a necessary and sufficient condition developed prev...
Preprint
Full-text available
Autonomous cyber-physical systems (CPS) rely on the correct operation of numerous components, with state-of-the-art methods relying on machine learning (ML) and artificial intelligence (AI) components in various stages of sensing and control. This paper develops methods for estimating the reachable set and verifying safety properties of dynamical s...
Conference Paper
Full-text available
Partial differential equations (PDEs) mathematically describe a wide range of phenomena such as fluid dynamics, or quantum mechanics. Although great achievements have been accomplished in the field of numerical methods for solving PDEs, from a safety verification (or falsification) perspective, methods are still needed to verify (or falsify) a syst...
Article
Full-text available
Reachability analysis is a fundamental problem for safety verification and falsification of Cyber-Physical Systems (CPS) whose dynamics follow physical laws usually represented as differential equations. In the last two decades, numerous reachability analysis methods and tools have been proposed for a common class of dynamics in CPS known as ordina...
Article
In this work we address the problem of (outer) estimation of reachable sets in switched linear systems subject to dwell-time switching. After giving some conditions that exploit the well-known properties of exponential decrease/bounded increase of the Lyapunov function (i.e. exponential decrease in between switching times and bounded increase at sw...
Article
Full-text available
In this work, the reachable set estimation and safety verification problems for a class of piecewise linear systems equipped with neural network controllers are addressed. The neural network is considered to consist of Rectified Linear Unit (ReLU) activation functions. A layer-by-layer approach is developed for the output reachable set computation...
Article
Full-text available
Neural networks have been widely used to solve complex real-world problems. Due to the complicate, nonlinear, non-convex nature of neural networks, formal safety guarantees for the behaviors of neural network systems will be crucial for their applications in safety-critical systems. In this paper, the reachable set estimation and verification probl...
Article
Partial differential equations (PDEs) mathematically describe a wide range of phenomena such as fluid dynamics, or quantum mechanics. Although great achievements have been accomplished in the field of numerical methods for solving PDEs, from a safety verification (or falsification) perspective, methods are still needed to verify (or falsify) a syst...
Article
Full-text available
Neural networks have been widely used to solve complex real-world problems. Due to the complicate, nonlinear, non-convex nature of neural networks, formal safety guarantees for the output behaviors of neural networks will be crucial for their applications in safety-critical systems.In this paper, the output reachable set computation and safety veri...
Article
In this paper, a novel parameter-memorized Lyapunov function is proposed for stability analysis of discrete-time linear systems with time-varying parametric uncertainties. The parameter-memorized Lyapunov function depends on the uncertain parameters with a memory of a certain interval. It is shown that the previous parameter-dependent Lyapunov func...
Article
Full-text available
In this note, the globally exponential stability of discrete-time switched systems under arbitrary switching is investigated. First, for discrete-time switched nonlinear systems, the globally exponential stability is found to be equivalent to the existence of an M-step sequence with sufficient length and a family of Lyapunov functions, and then a s...
Article
Full-text available
In this paper, the output reachable estimation and safety verification problems for multi-layer perceptron neural networks are addressed. First, a conception called maximum sensitivity in introduced and, for a class of multi-layer perceptrons whose activation functions are monotonic functions, the maximum sensitivity can be computed via solving con...
Article
This paper deals with the problems of stability analysis and L1-gain characterization for continuoustime switched systems consisting of positive subsystems. With the aid of a discretized copositive Lyapunov function, a sufficient condition ensuring the asymptotic stability of continuous-time switched positive systems under dwell-time constraint is...
Article
Full-text available
Order-reduction is a standard automated approximation technique for computer-aided design, analysis, and simulation of many classes of systems, from circuits to buildings. For a given system, these methods produce a reduced-order system where the dimension of the state-space is smaller, while attempting to preserve behaviors similar to those of the...
Article
Full-text available
This paper addresses the output reachable set estimation problem for continuous-time switched linear systems consisting of Hurwitz stable subsystems. Based on a common Lyapunov function approach, the output reachable set is estimated by a union of bounding ellipsoids. Then, multiple Lyapunov functions with time-scheduled structure are employed to e...
Article
Full-text available
The event-triggered control problem for switched linear system is addressed in this study. The periodical sampling scheme and event-triggering condition are incorporated in the closed-loop. The feedback control updates its value only at sampling instants as long as event-triggering condition is satisfied as well. In addition, the switchings are onl...
Article
In this note, a necessary and sufficient stability criterion for switched linear systems under dwell-time constraint is proposed by employing a class of time-scheduled homogeneous polynomial Lyapunov functions with a sufficiently large degree. The key feature of this nonconservative condition lies in its convexity in the system matrices, which expl...
Conference Paper
Full-text available
Hybrid systems have both continuous and discrete dynamics and are useful for modeling a variety of control systems, from air traffic control protocols to robotic maneuvers and beyond. Recently, numerous powerful and scalable tools for analyzing hybrid systems have emerged. Several of these tools implement automated formal methods for mathematically...
Article
In this paper, an elementary time unit (ETU) method is proposed to study the problem of estimating the admissible delay in the identification of the active mode in the analysis and design of switched systems. The activation interval of a subsystem is considered to consist of a finite-time number of segments called ETUs, by which a novel class of ti...
Article
In this paper, the dissipativity and dwell time specifications of switched discrete-time system are investigated. In general case, the dissipativity and the desirable induced Lyapunov stability of switched systems can be established by imposing non-increasing condition on multiple storage functions at switching instants. Then, a novel dwell-time de...
Article
In this note, a study on the equivalence of two stability criteria for continuous-time switched linear systems with dwell time constraint is presented. It is demonstrated that, for any dwell time satisfying the conditions in stability criterion proposed in Geromel and Colaneri (2006), the conditions in stability criterion of Allerhand and Shaked (2...
Article
In this paper, the stability and failure tolerant control are investigated for discrete-time time-invariant linear systems with occasional controller failure due to physical or purposeful reason. By introducing the conceptions of minimal controller reliable time and maximal controller failure time, a novel time-varying piecewise quadratic Lyapunov...
Article
Full-text available
In this technical note, the stability and failure tolerant control are investigated for linear time-invariant (LTI) control systems with occasional controller failures. At first, a collection of non-convex sufficient conditions ensuring the closed loop to be globally asymptotically stable is presented, which are simple to use but difficult to consi...
Article
In this paper, the control synthesis problem for a class of large-scale systems with multi-modes that are called large-scale switched systems is addressed. By introducing the concept of decentralized switching signal and the relevant decentralized average dwell time, the asymptotic stability and weighted ℓ2 gain performance are investigated. It sho...
Article
In this paper, the reliable tracking control problem for high speed trains against actuator failures is addressed. A multiple-point-mass model for the high speed train against actuator failures is presented, in which the actuator failures including both partial loss of actuator effectiveness and complete breakdown are characterized by unknown varyi...
Article
In this paper, the traffic-responsive signalization problem for an oversaturated intersection is addressed. A static state feedback control strategy is presented to relieve the waiting vehicles in lanes. By modeling the intersection as a discrete-time switched system and with the aid of quadratic Lyapunov function method, the green time of each pha...
Article
In this article, the filtering problem for switched discrete-time linear systems under asynchronous switching is addressed in the framework of dwell time, where ‘asynchronous switching’ covers more general and practical cases, for example, the switching lags caused by mode identification process are taken into consideration. Firstly, a novel dwell-...
Article
In this technical note, the l(2) gain analysis and asynchronous H-infinity control problems for switched discrete-time linear systems are addressed. Under the dwell time constraint, the computation of l(2) gains versus dwell time is viewed as an open problem. By introducing a novel dwell time dependent Lyapunov function, an LMI-based sufficient con...
Article
The Luenberger observer technique has long been used for reliable state estimation of feedback control systems. In this paper it is employed for asynchronously switched control systems. The theory of finite-time stability (FTS) and finite-time boundedness (FTB) is incorporated, which is a relatively new concept offering open and challenging problem...
Article
In this paper, a new type of switched parallel distributed compensation (PDC) controller is presented, in which a novel switching scheme, namely, membership function ranking (MFR) scheme is proposed. The MFR scheme can be viewed as an improvement and extension of previous results named largest membership function (LMF) scheme concerned with piece-w...
Article
In this note, the state estimation problem for a class of short-time switched linear systems is investigated. By incorporating the concept of finite-time stability, an observer is designed to ensure the error dynamics finite-time stable in the short-time switching interval under the general asynchronous switching whereas synchronous switching mode...
Article
Stabilization of switched systems composed fully of unstable subsystems is one of the most challenging problems in the field of switched systems. In this brief paper, a sufficient condition ensuring the asymptotic stability of switched continuous-time systems with all modes unstable is proposed. The main idea is to exploit the stabilization propert...
Article
In this paper, a new approach for stability analysis of time-dependent switched linear systems is proposed. System equivalence is the main idea in this new approach, which derives a switched discrete linear parameter-varying system from the switched continuous-time linear switched system with interval dwell time, and the stability properties of the...
Article
Full-text available
In this brief article, the stability and disturbance attenuation property in the sense of ℓ2 gain for switched linear discrete-time systems are investigated. The computation of ℓ2 gains against dwell time has been viewed as an open problem so far. A novel dwell-time dependent Lyapunov function (DTDLF) is introduced to study the stability and ℓ2 gai...
Article
In many actual engineering applications, the Markovian jump among subsystems often occurs in some short finite-time intervals. For this class of Markovian jump systems, called short-time Markovian jump systems in this paper, the state boundedness during the short-time Markovian jump interval is of great interest. By introducing the concepts of fini...
Article
Because switched system with nonlinear subsystems shows great difficulties in controller design by conventional methodologies, the fuzzy logic theory, which has been proven to be a practical and effective way to deal with the analysis and synthesis problems for complex nonlinear systems, is applied to design a class of hybrid fuzzy controller for s...
Article
Full-text available
The concept of finite-time stability (FTS) has gained much attention from researchers in recent years, after the emergent utilization of linear matrix inequalities (LMIs) in the field of control systems. In this paper, FTS and finite-time boundedness (FTB) analysis and synthesis procedure for a ‘finite-time H∞ filter’ for discrete-time switching sy...
Article
Fuzzy T-S model has been proven to be a practical and effective way to deal with the analysis and synthesis problems for complex nonlinear systems. As for switched nonlinear system, describing its subsystems as fuzzy T-S models, namely switched fuzzy system, naturally is an alternative method to conventional control approaches. In this paper, the H...
Article
The H ∞ filtering problem of switched non-linear systems with polytopic uncertainties is addressed in this paper. A novel filter with state impulsive jump is proposed to ensure the stability of error system and guarantee a prescribed noise attenuation level in the H ∞ sense by solving linear matrix inequalities (LMIs) and determining an impulsive j...
Article
Mathematical models are an approximate of physical systems and design procedures are only complete when modeling errors have been quantified. Uncertainties are incorporated in design procedure to compensate such discrepancies and to add robustness. This paper investigates the design problem of parameter-dependent switched observers for polytopic un...
Article
In this paper, the fault detection problem for a class of state-dependent switched nonlinear systems with linear switching surface is addressed. The investigation of fault detection problem includes two parts: design sub-filters for each subsystem, and determine a proper update of estimated state. A fault detection filter is proposed incorporating...
Article
This paper concerns the state estimation problem for a class of uncertain non-linear Markovian jump systems, where the Markovian jump only occurs in some ‘short’ finite-time intervals. For this class of Markovian jump systems, the boundedness of estimation error deserves our investigation. By introducing the concepts of finite-time stochastic stabi...
Article
Full-text available
In this paper, the H filtering problem for a class of short-time switched discrete-time linear systems is investigated. For such systems, switching always occurs in some short interval. Since the error state may attain large unacceptable values in short-time switching intervals, besides the asymptotic stability of error dynamics, the boundedness of...
Article
Full-text available
This note addresses a defect occurring in [1], in which the desired l(2)-gain should be replaced by weighted l(2)-gain performance. The corrected results and proof are presented.
Article
In this article, a new parameter-dependent stability criterion for switched linear discrete-time systems with polytopic uncertainties is provided to correct an error occurring in previous work, and the parameter-dependent idea is extended to l(2) gain analysis problem. Particularly, a less conservative result is proposed by introducing additional s...
Article
In this paper, we investigate the stability analysis problem of switched impulsive nonlinear systems and several stabilization problems of switched discrete-time linear systems are studied. First, sufficient conditions ensuring globally uniformly asymptotically stability of switched nonlinear impulsive system under arbitrary and DDT (dynamical dwel...
Article
In this paper, several concepts of switching frequency are introduced to analyze the properties and performance of switched systems in infinite as well as finite-time intervals. The observation is very motivating that different system properties and performances depend on different switching frequencies. Sufficient conditions ensuring asymptotic st...