Biomedical Signal Processing and Control 49 (2019) 375–387 Contents lists available at ScienceDirect Biomedical Signal Processing and Control journal homepage: www.elsevier.com/locate/bspc A nonovershooting tracking controller for simultaneous infusion of anesthetics and analgesics  Regina Padmanabhan a , Nader Meskin a,∗ , Clara M. Ionescu b,c,d , Wassim M. Haddad e a Department of Electrical Engineering, Qatar University, Qatar b Department of Electrical Energy, Ghent University, Technologiepark 913, 9052 Gent-Zwijnaarde, Belgium c Technical University of Cluj Napoca, Memorandumului Street No. 28, Cluj, Romania d Flanders Make, core lab EEDT, Belgium e School of Aerospace Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0150, USA a r t i c l e i n f o Article history: Received 8 April 2018 Received in revised form 27 August 2018 Accepted 27 September 2018 Keywords: Nonovershooting control Active drug dosing Biomedical control a b s t r a c t In this paper, a nonovershooting tracking controller is proposed for the continuous infusion of multi- ple drugs that have interactive effects. The proposed controller design method exploits the freedom of eigenstructure assignment pertinent to the design of feedback controllers for multi-input, multi-output (MIMO) systems. For drug dosing, a nonovershooting tracking controller restricts the undesirable side effects of drug overdosing. The proposed tracking controller is based on an estimate of the full state using a hybrid extended Kalman filter (EKF) that is used to reconstruct the system states from the measurable system outputs. To illustrate the proposed method, we use one of the common anesthetic and analgesic drug combinations (i.e., propofol and remifentanil) which exhibit nonlinear and synergistic drug inter- action. An integral control action is included in the controller design to achieve robust tracking in the presence of patient parameter uncertainty. Simulation results and performance analysis of the proposed control strategy are also presented using 20 simulated patients. © 2018 Elsevier Ltd. All rights reserved. 1. Introduction Critically ill patients in the intensive care units (ICUs) often require fine tuned and long term infusion of anesthetics and analgesics [1]. The term analgesia denotes blunting of pain by med- ication. In ICUs, sedative and analgesic drugs are used to reduce anxiety, delirium, decrease pain during intubation and extubation, increase patient tolerance after endotracheal tube insertion, and to reduce patient ventilator dysynchrony. Even though the critical task of anesthesia administration has been widely discussed in the literature and studied using clinical trials over the last few decades, several recent reviews on the existing methods highlight various aspects of the problem that needs further research attention [2,3]. Such attempts can help transform the research knowledge in this area into clinically and educationally applicable (bedside) tools [2].  This publication was made possible by the GSRA Grant No. GSRA1-1-1128-13016 from the Qatar National Research Fund (a member of Qatar Foundation). The findings achieved herein are solely the responsibility of the authors. ∗ Corresponding author. E-mail addresses: regina.ajith@qu.edu.qa (R. Padmanabhan), nader.meskin@qu.edu.qa (N. Meskin), ClaraMihaela.Ionescu@UGent.be (C.M. Ionescu), wm.haddad@aerospace.gatech.edu (W.M. Haddad). In the case of combined administration of anesthetic and anal- gesic drugs, the mechanism of action is complex, interlaced, and not yet completely understood making the problem more challenging. There are several factors that influence the pharmacokinetics and pharmacodynamics of a drug in the human body. Patient param- eters such as the height, weight, gender, age of the patient, and illnesses associated with the circulation system, renal system, or respiratory system influence the pharmacokinetics and pharma- codynamics of a drug in the patient’s body [4–6]. For continuous infusion of anesthetic and analgesic drugs over long periods, it is apparent that an appropriate closed-loop control strategy can be used to enhance patient safety [1,3,5]. Two basic factors to consider while designing a closed-loop controller for anesthesia administration are choosing appropriate parameters for feedback and deciding on a viable control strategy to implement. We use a common sedation assessment measure such as the bispectral index (BIS) to quantify the sedation level of the patient [7,8]. However, when it comes to the assessment of pain levels, a relatively smaller number of measurable variables are identified to quantify the pain levels and thus, facilitate closed-loop control [1,3]. In ICUs, several analgesic drugs are used to suppress the pain sensation and thereby reduce stress. The most reliable and valid indicator of pain is the patient’s self-report [1]. However, critically https://doi.org/10.1016/j.bspc.2018.09.015 1746-8094/© 2018 Elsevier Ltd. All rights reserved.