Article Transactions of the Institute of Measurement and Control 1–14 Ó The Author(s) 2019 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/0142331219864188 journals.sagepub.com/home/tim Integral backstepping-based output feedback controller for the induction motor Soufien Hajji 1 , Assil Ayadi 1 , Youssef Agerbi Zorgani 1 , Tarak Maatoug 1 , Mondher Farza 2 and Mohamed M’Saad 2 Abstract This paper addresses the control of induction motor (IM) drives. In this work, we propose a new consideration of backstepping control. However, this control provides a systematic method to carry out the controller design while guaranteeing the stability of the controller-process couple. Furthermore, the incorporation of an integral action in the synthesis of the control system with state feedback presents a robust rejection of echelon-level distur- bances. A detailed analytic study and simulation results are given showing the operation of the IM drives control. The results prove the accuracy and robustness of the proposed control scheme. Also, comparison results with another study dealing with control prove that the proposed method shows excellent transient and steady-state speed and a great estimation of flux and load torque. Keywords Nonlinear system, output feedback control, high gain control, backstepping control, high gain observer, integral action, induction motor Introduction Induction motor (IM) drives are more rugged, compact, cheap and reliable in comparison with the other motors used for industrial applications (Bose, 2002; Delaleau et al., 2001; Leonard, 2001; Lubineau et al., 2000). Furthermore, nowa- days, the field of variable speed drive has been characterized by many actuators associating AC machines and static inver- ters (Lubineau et al., 2000; Strangas et al., 1999). In the past, this solution was not possible, principally due to the complex structures of this type machine control (Blaschke, 1971). Their multi-variable and non-linear models, non-measurable states and variation of parameters during operation have limited these performances. However, the increasingly rapid progres- sion of digital processors has enabled the implementation of sophisticated control techniques to achieve high performance in terms of speed, stability and precision (Ghanes and Zheng, 2009; Ghanes et al., 2010). Because of the powerful computers development such as the microcontrollers, the DSPs, and so forth, which facilitate the implementation at low cost of the complex real-time control algorithms in the current industrial computing systems (Ghanes and Glumineau, 2010; Lubineau et al., 2000). Without forgetting the technological evolution of the semiconductors that constitute the static inverters associ- ated with and without the control systems, it is possible for a long time to operate the asynchronous machine at variable frequency. The power supply and calculations difficulties being adjusted, however many control strategies have been developed. Thus, they permit to widen the power range used by the asynchronous machine in the speed variation controls. Numerous research works have been presented to clarify the control of IM, such as the field oriented control (FOC) (Tajima and Hori, 1993; Traor et al., 2007), the direct torque control (DTC) (Soreshjani, 2015; Strangas et al., 1999), the sliding mode control (SMC) (Mahmoud et al., 2006), the high gain control (HGC) (Hajji et al., 2008) and backstepping con- trol (BC) (Ouadi et al., 2002; Tan and Chang, 1999). Moreover, in the literature other recent controllers are pro- posed for a class of nonlinear system such as network control- ler design (Li et al., 2019; Liu et al., 2018a,b), and fuzzy based multi-error constraint control (Gao et al., 2018). The FOC was introduced by Brown, Kovacs and Vas (1983) and revival by Blaschke (1971). Indeed, this method requires a very high computation time such as the calculation of the Park function, the vector processing. Therefore, this control could be implemented only after having carried out progress on the microprocessors. Thanks to this evolution, the central processing unit (CPU) time of the loops became negligible (Kan et al., 2015). The disadvantage of this method is its sensitivity to the parametric variations. In 1986, the 1 Laboratory of Sciences and Techniques of Automatic Control and Computer Engineering (Lab-STA), National School of Engineering of Sfax, University of Sfax, Tunisia 2 Laboratoire d’Automatique de Caen, ENSICAEN, Universite´ de Caen, France Corresponding author: Soufien Hajji, Laboratory of Sciences and Techniques of Automatic Control and Computer Engineering (Lab-STA), National School of Engineering of Sfax, University of Sfax Postal Box 1173, 3038 Sfax, Tunisia. Email: hjjisfin@yahoo.fr