International Journal of Dynamics and Control https://doi.org/10.1007/s40435-021-00878-1 Swarm-based robust fixed-structure controller design for buck converter using Kharitonov approach: design and experiment Ali Ghassab Sedehi 1 · Alireza Alfi 1 Received: 4 June 2021 / Revised: 20 August 2021 / Accepted: 19 September 2021 © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2021 Abstract A key challenge of DC–DC converters is designing an appropriate controller for reaching the output voltage to steady-state in a limited time with small variation. In addition, the parameters of converters may be affected by different factors that can make difficulties to voltage regulation. Robust control theory is a significant method to deal with this problem. Despite high efficiency, the high-order robust controllers may not be feasible for real-time implementation due to the hardware and computational limitations. In this paper, we study the swarm-based robust controller design with fixed-structure for DC–DC converter using the Kharitonov approach. The developed algorithm allows implementing the simple structure controllers that guarantees both robust stability and performance. The controller behavior is validated and compared with other related works. Experiments are provided to demonstrate the feasibility of the designed controller. Keywords Buck converter · Kharitonov approach · Fixed structure · Robust control · Optimization · Swarm 1 Introduction DC–DC Buck converter is a type of step-down converter, which has been widely applied in different fields to supply a fixed amount of DC voltage. Its simple structure, low cost, and lightweight make it have a key place in connections to smartphones, wind turbines, storage batteries, robotics, electromotive automobiles, DC microgrids, and photovoltaic (PV) systems [1–5]. Buck converter provides a challenging field in designing the controller due to its switching operation, time-varying behavior, and inherent uncertainty and distur- bances. Consequently, linear conventional control methods may not ensure a broad operation range [6]. A large body of researches has been studied using different control tech- niques like fuzzy logic control [7–11], sliding mode control techniques [12–18], adaptive backstepping control [19–21], feedback linearization control [22], fractional control [23], and linear-quadratic Regulator (LQR) [24]. B Alireza Alfi a_alfi@shahroodut.ac.ir Ali Ghassab Sedehi ali.sedehi@yahoo.com 1 Faculty of Electrical Engineering, Shahrood University of Technology, Shahrood, Iran In practice, the system performance is frequently influ- enced by the parameters perturbations, resulting in the uncer- tainties [25]. Consequently, the most significant purposes in regulation are to ensure both stability and performance requirements against the uncertainties [26–28]. Robust con- trol is one of the well-known techniques to achieve these goals. A key technique for designing robust controllers is the μ-synthesis [29], which requires a linearized model of the system and can include several weighting functions for shaping the exogenous signals and representing the sys- tem’s performance specifications. According to literature, the results confirmed that this method provided much bet- ter results than the conventional classical controllers [30]. Generally, μ-synthesis controller conduces to high order controllers, which is a major disadvantage that makes the controller may not realizable in practice due to the limita- tions of hardware and computational burden. This problem plays an important role to possibility of implementation of the resultant control laws [31]. The D-K iteration algorithm is the most well-known approach for solving μ-synthesis prob- lem suffers from a major disadvantage that cannot converge in some cases, resulting in the non-optimality of the designed controller [32]. The systematic method is H ∞ control frame- work that the model order reduction of the controller can be adopted leading to degradation of the system performance and robustness as the H ∞ norm increases [33]. Another 123