Ph. D. Project
Title:
Fault and attack detection and identification for cyber-physical systems
Dates:
2019/10/01 - 2022/09/30
Description:
Context
Cyber-physical systems are characterized by the integration of physical processes and computing and communication capabilities. For example, such systems can be found in various fluid transportation and distribution networks (gas, electricity, water, etc.). These systems are said to be cyber-physical since they encompass a physical part (pipelines and drinking water tanks, pumping stations, for example) but also include sensing devices to measure physical growths (flow and pressure sensors in water network, physical and chemical measures of the water quality), information transmission networks, as well as control devices using the measured quantities to act on the physical part. These systems, in addition to their own potential failures (component degradation, information transmission faults, for example), have to deal with malicious external attacks that can severely degrade their operating. The monitoring of cyber-physical systems is therefore a major issue : it is essential to be able to detect, locate and identify these degradations and / or external attacks.

Expected researches
In order to take into account the material or information conservation laws in physical or communication networks, models of cyber-physical systems involve both dynamic models and static relationships. Modeling in descriptor form (one also speaks of algebro-differential or singular systems) is then natural (Xu, 2006 ; Yang, 2013 ; Lopez Estrada, 2014 ; Estrada Manzo, 2015). In the case of linear models, tools borrowed from the graph theory and automatic control theory respectively allow a structural analysis of attacks (detectability, distinguishability, etc.) and their estimation (Pasqualetti, 2011 ; Pasqualetti, 2013). The purpose of the thesis work would be to extend this approach to the more general nonlinear framework, thus allowing the application of these results to a larger number of cyber-physical systems or allowing this application to be more accurate by avoiding use models that are too simple and therefore too approximative to represent a complex reality. For this, the polytopic or LPV approach will be used to efficiently model nonlinear phenomena while benefiting from certain advantages of linear structures (Lendek, 2010, Tanaka, 2001).

Attacks and faults to be considered in this thesis are as follows.
— External attacks can take the form of measurement corruptions affecting the cyber-physical systems : unknown inputs substituting or corrupting the transmitted data (Pasqualetti, 2011) ;
— Degradations can be represented by transmission faults from or to the physical part : missing data (Zhao, 2009), saturations (Tarbouriech, 2011), dead zones (Chiang, 2016) ;
— It is also necessary to take into account the faults of the physical part : degradation of the system himself (wear).

After a bibliographic study, the solutions that the candidate might consider in the context of of this thesis are as follows :
— descriptor polytopic / LPV model based diagnosis, based on observers ;
— descriptor polytopic / LPV model based diagnosis by coprime factorization, this technique has been used in the linear framework for diagnosis (Frank, 1994) and for fault tolerant control (Zhou, 2001), but its extension to the polytopic framework remains open;
— synthesis of diagnosis filters for descriptor polytopic / LPV models;
— polytopic modeling of the phenomena of saturation and / or dead zones allowing their taking into account in the model of the system, even their estimation (Bezzaoucha, 2016).

References
(Bezzaoucha, 2016) S. Bezzaoucha, B. Marx, D. Maquin, J. Ragot, State and output feedback control for Takagi-Sugeno systems with saturated actuators, International Journal of Adaptive Control and Signal Processing, 30 : 888-905, 2016.
(Chiang, 2016) C.C. Chiang, C.T. Tsai, Model Reference Fuzzy Sliding Mode Control of Strict- Feedback Uncertain Systems with Unknown Dead-Zone, IEEE International Conference on Fuzzy Systems, 2016.
(Estrada Manzo) V. Estrada Manzo, Estimation et commande des systèmes descripteurs, Thèse de doctorat de l'Université de Valenciennes et du Hainaut-Cambrésis, 2015.
(Lendek, 2010) Z. Lendek, T.M. Guerra, R. Babuska, and B. De Schutter. Stability Analysis and Nonlinear Observer Design using Takagi-Sugeno Fuzzy Models, Springer, 2010.
(Lopez Estrada, 2014) R. Lopez Estrada, Contribution au diagnostic de défauts à base de modèles : Synthèse d'observateurs pour les systèmes singuliers linéaires à paramètres variants aux fonctions d'ordonnancement non mesurables, Thèse de doctorat de l'Université de Lorraine, 2014.
(Pasqualetti, 2011) F. Pasqualetti, F. Dorfler, F. Bullo, Cyber-physical attacks in power networks: Models, fundamental limitations and monitor design, IEEE Conference on Decision and Control, 2011.
(Pasqualetti, 2013) F. Pasqualetti, F. Dorfler, F. Bullo, Attack Detection and Identification in Cyber-Physical Systems, IEEE Transactions on Automatic Control, 58(11) :2715-2729, 2013.
(Tanaka, 2001) K. Tanaka and H.O.Wang. Fuzzy Control Systems Design and Analysis : A Linear Matrix Inequality Approach. Wiley, 2001.
(Tarbouriech, 2011) S. Tarbouriech, G. Garcia, J.M. Gomes da Silva, Jr., and I. Queinnec. Stability and Stabilization of Linear Systems with Saturating Actuators. Springer, 2011.
(Xu, 2006) S. Xu and J. Lam. Robust Control and Filtering of Singular Systems. Lecture Notes in Control and Information Sciences. Springer, 2006.
(Yang, 2013) C. Yang, Q. Zhang, and L. Zhou. Stability Analysis and Design for Nonlinear Singular Systems. Springer, 2013.
(Zhao, 2009) Y. Zhao, J. Lam, H. Gao, Fault Detection for Fuzzy Systems With Intermittent Measurements, IEEE Transactions on Fuzzy Systems, 17(2) : 398-410, 2009.
(Zhou, 2001) K. Zhou and Z. Ren. A new controller architecture for high performance, robust and fault-tolerant control. IEEE Transactions on Automatic Control, 46(10) :1613-1618, 2001.
Keywords:
Cyber-physical systems, Attack detection, Fault diagnosis
Conditions:
Faculty : Université de Lorraine
Laboratory : CRAN Research Center for Automatic Control of Nancy (website)
PhD advisors : Benoît Marx (website) and Jean-Christophe Ponsart.
Beginning of the PhD : 09/01/2019
Duration of the thesis : 3 years
Grant : LUE IMPACT DigiTrust "Sécurité du citoyen dans le monde numérique"
Salary : approx. 1500 euros per month
Candidate's profile : M.Sc in Automatic Control
Contact : benoit.marx@univ-lorraine.fr and jean-christophe.ponsart@univ-lorraine.fr
Department(s): 
Control Identification Diagnosis
Funds:
LUE IMPACT DigiTrust "Sécurité du citoyen dans le monde numérique", Contrat Doctoral AM2I (demande)