A cooperative fault-tolerant control method for the coupled system based on interaction effect utilization Jing Chang ∗ J´ erˆ ome Cieslak ∗∗ Zongyi Guo ∗∗∗ David Henry ∗∗ ∗ Xidian University, Xi’an, China (e-mail: jchang@xidian.edu.cn). ∗∗ University of Bordeaux, Talence 33400, France (e-mail: jerome.cieslak@ims-bordeaux.fr;david.henry@u-bordeaux.fr) ∗∗∗ Northwestern Polytechnical University, Xi’an, China (e-mail: guozongyi@nwpu.edu.cn) Abstract: An active fault-tolerant control scheme is proposed in this paper for the strongly coupled MIMO systems which subject to total actuator failures. A control-oriented interaction indicator is defined in the Lyapunov stability sense and then is utilized to design the cooperative fault-tolerant control law. The proposed scheme can achieve robust tracking performance with globally uniformly ultimate boundlessness and is capable of improving transient performance (fault-tolerant ability) by wisely using the interactions. Simulation results obtained on a flight attitude control system illustrates the benefit of the proposed techniques. Keywords: Fault tolerant control, cooperative control, interaction effect utilization. 1. INTRODUCTION With the rapid development of unmanned autonomous systems, the past several decades have witnessed an ex- plosive growth of Fault Detection and Diagnosis (FDD), Fault Detection and Isolation (FDI), Fault-Tolerant Con- trol (FTC) and Fault-Tolerant Guidance (FTG) in most disciplines of engineering to enhance the safety of the system in case of the occurrence of unknown fault/failure, see Zolghadri et al. (2014). And particularly, safety is a critical issue in the area of aircraft and aerospace indus- try, where the FDI, fault-tolerant flight control have been widely developed (Gao and Wang (2014)). Reliable control systems are needed more than ever in the face of rising autonomous and intelligent levels, increasing advancement and complexity of aircraft and aerospace vehicles, the op- portunities and challenges presented by new technologies (distributed, networked and cooperative). Despite the high number of published works about fault-tolerant control, too few works take into account the total actuator failures. Control Allocation (CA) is an efficient approach for deal- ing with total actuator failures without the reconfiguration of the controller (Alwi and Edwards (2008)). However, it is hard to consider the uncertain control effectiveness and obtain the optimized CA.On the other hand, the CA method required the invertible of the control distribution matrix, which is not always held for the serious actuator faults. As for a faulty system within total actuator failures, it sometimes becomes an underactuated system. The exis- tence of state interactions and control interactions in the MIMO system is a form of coupling. The strong coupling exists in some MIMO systems (see Weiland et al. (2004); ⋆ This work was supported by the National Natural Science Foun- dation of China under Grant 61803308. Corresponding author Jing Chang (jchang@xidian.edu.cn). Tian et al. (2013)) which results in difficulty to control the system in fast maneuvering motions, such as rapid descent and fast lateral maneuver of flight. For the normal case, the decoupling control scheme (Dhadekar and Patre (2017)) and decentralized control scheme (Dickeson et al. (2009)) can help in easily and directly designing the controller, since it allows the designer to set the controller for each state independent of the other states. When applicable, the advantage of completely decentralized control is that one can apply the simpler SISO theory (Dhadekar and Patre (2017)). While this traditional design scheme essentially regards the coupling as a detrimental element in the sys- tem and eliminates or suppresses its effect on the system directly (Guo et al. (2017, 2018a)). In fact, the existence of interactions between the states or inputs of MIMO systems could be helpful to stabilize the actual motion in under-actuated systems. As for couplings acting on the MIMO systems, the fault-tolerant property of interactive effect has not been sufficiently investigated yet, nor the relationship with the expected system trajectory. In recent work (Guo et al. (2017, 2018b); Chang et al. (2018)), a novel coupling effect indicator is proposed to demonstrate the coupling effects on the system. This proposed con- trol scheme achieved a better dynamic performance by explicitly utilizing the system couplings in the controller design. The improvements in the dynamic performance of this strategy thanks to the switching behavior triggered by the coupling effect. This control strategy is similar to the phase-based gain-modulation control that improves damping while the error is increasing and reducing the control gain while transiting toward the desired output. This motivates the work of this paper, wherein the con- troller is modulated according to interaction indicators to attain a certain cooperative fault-tolerant method for MIMO systems. More specifically, this paper will revisit Preprints of the 21st IFAC World Congress (Virtual) Berlin, Germany, July 12-17, 2020 Copyright lies with the authors 4175