Olaf Simanski

The benefits of using Guyton's model in a hypotensive control system

In order to improve the intraoperative applications, this paper presents the advantages of using Guytons model in hypotensive control system development. In this system, the mean arterial pressure is decreased and maintained at a low level during anaesthesia by controlling sodium nitroprusside infusion rate. The key of the study is to develop a physiological model of cardiovascular dynamics to present the mean arterial pressure response to sodium nitroprusside, which was considered as a linear model in most of known blood pressure control systems. Being linear, the previous models cannot accurately mimic a physiological system of human circulation, especially at deep hypotensive control with strong reaction of the body. The enhanced model in this study was modified based on Guytons model of human circulation. It is useful to design a PID controller, which allows studying and handling the wide range of the body sensitivities. This model is also helpful for studying the behaviors of patients under anaesthesia conditions, such as the perfusion of organs and the reaction of the body at hypotensive state. A fuzzy gain scheduler and a supervising algorithm were also developed for online tuning the controller to handle the behavior of the body. The control system was tested on 25 experiments on seven pigs in the animal laboratory. Simulation and experiment results proved the usefulness of Guytons model in control system design which can present the dynamical response of blood pressure in the circulation under and after hypotensive control. The results also indicated the safety and stability of the controller.

Design and implementation of a control system reflecting the level of analgesia during general anesthesia.

Abstract Introduction: Measuring and ensuring an adequate level of analgesia in patients are of increasing interest in the area of automated drug delivery during general anesthesia. Therefore, the aim of this investigation was to develop a control system that may reflect the intraoperative analgesia value. Our hypothesis was that a feedback controller could be applied in clinical practice safely and at an adequate quality of analgesia. The purpose of this study was to evaluate the practical feasibility of such a system in a clinical setting. Methods: The control system for the level of analgesia described in this paper relies on a parameter combination of heart rate variability (HRV), heart rate (HR), and blood pressure (mean arterial pressure, MAP), which serve as input variables for an expert system. For this fuzzy system, the experience of the participating anesthesiologists was translated into a set of fuzzy rules. In a pilot trial, the control system for automated titration of remifentanil, a short-acting opioid, was tested combined with a closed-loop propofol infusion system for hypnosis. Ten adult patients (4 women, 6 men), aged 22–52 years (median, 45 years; range, 29–49 years), with an American Society of Anesthesiologists physical status class I or II and who were scheduled for elective trauma surgery in a supine position were enrolled in this prospective trial. The precision of the system was calculated using internationally defined performance parameters. Results: There was no human intervention necessary during the computer-controlled administration of propofol and remifentanil, and operating conditions were satisfactory in all patients. All patients assessed the quality of anesthesia as “good” to “very good”. Median performance error, median absolute performance error, and wobble for HR and MAP during maintenance of anesthesia were -8.98 (5.32), 10.08 (4.17), and 2.68 (1.29) and -4.51 (12.73), 13.63 (2.27), and 3.90 (2.08) [mean (SD)], respectively. Conclusion: The control system, reflecting the level of analgesia during general anesthesia designed and evaluated in this study, allows for a clinically practical, nearly fully automated infusion of an opioid during medium-length surgical procedures with acceptable technical requirements and an adequate precision.

Automatic drug delivery in anesthesia: From the beginning until now

The main goals of general anesthesia are adequate hypnosis, analgesia and maintenance of vital functions. For some special kinds of operations neuromuscular block is essential. Furthermore the patient safety and cost reduction as a case of minimized drug consumption and shortened postoperative recovery phases are part of the main issues to motivate automation in anesthesia. Since the beginning of the eighties engineers and physician are working together in the field of the development of closed-loop systems for drug delivery. The work gives a short overview about the development of the automation in drug delivery systems over the last years without the claim of completeness and expressed the much more vision.

An online fuzzy gain scheduling for blood pressure regulation

Abstract This paper presents a method of fuzzy gain scheduling for a PID controller applied to mean arterial pressure (MAP) regulation during general anesthesia by sodium nitroprusside (SNP) infusion. A supervising algorithm is used for online updating the fuzzy gain scheduler of the PID controller to act stronger against the body reaction. A new model based on Slates model and Furutanis ideas was developed for testing the controller. Simulation and clinical results in deep hypotensive controls at 40mmHg on pigs indicate the safety and stability of designed controller.

Automatic drug delivery in anesthesia - the design of an anesthesia assistant system ⋆

Abstract The main goals of general anesthesia are adequate hypnosis, analgesia and maintenance of vital functions. For a number of surgical procedures neuromuscular block is essential. Furthermore patient safety and cost reduction as minimized drug consumption and shortened post-operative recovery are main issues and motivation of automation efforts in anesthesia. Since the beginning of the eighties engineers and physician are working together in the field of the development of closed-loop systems for drug delivery. The work gives only a short overview about the development of the automation in drug delivery systems over the last years without the claim of completeness and expressed the much more vision. In the final stage, the designed control system, the so called “Rostocker assistant system for anesthesia control (RAN)” should be fitted out with the possibility to control four different drugs automatically. Currently the multiple-input multiple-output (MIMO) control of the depth of hypnosis and neuromuscular blockade is realized as well as the closed-loop control of deep hypotension. A pilot study for the closed-loop control of analgesia is currently running. The paper points some general remarks and the designed MIMO-system for controlling the depth of hypnosis and the neuromuscular blockade

Monitoring the stress response during general anaesthesia

The major objectives of anaesthetists in the operating theatre are to maintain the hypnosis, the relaxation, the analgesia, and the vital functions during the operation. The aim of this paper is to present an overview of new developments regarding monitoring and control of stress response during general anaesthesia. The applicability of the main parameters to monitor the adequacy of analgesia online will be discussed. Different models based on pharmacokinetic and pharmacodynamic effects in particular the opioid remifentanil and the hypnotic propofol were developed. The response of these anaesthetics on heart rate, different parameters of the heart rate variability and the bispectral index during induction of general anaesthesia were studied. Starting from these models, control strategies in analgesia have been discussed. The developed environment for the assistant system for anaesthesia at the University of Rostock is presented.

Online Process Identification of Hypnosis

For most kinds of surgeries an adequate anaesthesia is necessary. Three main components characterise an anaesthesia - the relaxation, the analgesia and the hypnosis. This paper treats the topic of the Propofol/ Remifentanil induced hypnosis. We try to find a suitable model structure for the online identification of this process during a surgery. After a review of known models from literature an own model approach was developed and validated. By an offline optimisation study the parameter estimation problem of the defined cost function was examined. An online approach based on a recursive, iterative gradient algorithm was implemented. The quality of the identified parameters could be verified using measurements of surgeries.

A fuzzy system for regulation of the analgesic remifentanil during general anaesthesia

During the last decade, the research group Anaesthesia Control at the University of Rostock (Germany) developed an assistant system to support anaesthetists in the operating theatre. The main objectives during general anaesthesia are appropriate hypnosis, analgesia, relaxation, and stable vital functions. The development of our controllers for hypnosis and relaxation were completed in the last years. The systems were established in the clinical routine in the university hospital. Until now there exists no standard method to measure the state of analgesia or stress response of the patient. In this paper a fuzzy approach is presented to control the application of the analgesic drug remifentanil based on the expert knowledge of anaesthetists. The anaesthetists monitor vital parameters for example arterial blood pressure and heart rate to apply the analgesic drug. Additionally, the heart rate variability was involved as input for the fuzzy analgesia system, which is realized with two controllers, one active at a time. The main controller regulates the continuous rate of remifentanil based on the change in the mean arterial blood pressure, the heart rate and its standard deviation. The second controller calculates a bolus amount of remifentanil based on the absolute values of input parameters, if the heart rate is higher than 90 bpm. Both fuzzy controllers are Mamdani inference systems. The usual time regime for general anaesthesia as well as the minimal and maximal dosage of remifentanil are regarded in the system. The development options of the systems and the first results by controller simulations are discussed.

Modellbasierte Regelung pneumatisch betriebener Herzunterstützungssysteme

Zusammenfassung Dieser Beitrag stellt neue Ansätze für die Regelung eines extrakorporalen Herzunterstützungssystems vor. Die Grundlage ist eine kaskadierte Regelung von zwei Teilsystemen. Im ersten Teilsystem finden nichtlineare modellbasierte Regelungsansätze Anwendung. Im äußeren Regelkreis erfolgt die Regelung auf Basis einer Modellprädiktion der Regelgröße, da keine Messwerte zur Verfügung stehen. Die vorgestellte Regelung des Pumpvolumens adaptiert sich an zeitvariante, nicht direkt messbare Patientenparameter. Abstract The paper presents new approaches for the control of an extracorporeal ventricular assist device. It is based on a cascaded control of two subsystems. In the first subsystem a non-linear mode-based control approach was applied. In the outer loop, the control is based on a model prediction of the controlled variable, since no measurements are available. The proposed control of the pump volume adapts to time-variant, not directly measurable patient parameters.

Application of Iterative Learning Control for Repetitive Processes in Breathing Therapy

Abstract The continuous positive airway pressure (CPAP) therapy is the gold standard to treat the obstructive sleep apnoea syndrome. This continuous pressure splints the upper respiratory tracts and prevents a collapse of the pharynx. If not being treated a collapse of the upper respiratory tracts reduces the blood oxygen concentration and increases the CO2 concentration leading to repeated sleep disruptions. Many CPAP-devices cannot maintain the reference pressure, which can lead to an uncomfortable perception for the patient. To reduce the pressure deviation in the mask we want to apply the Iterative Learning Control (ILC) method, because the breathing of the patient is a repetitive process - and at night the breathing patterns are even and recurring from breath to breath.