Amélie Chevalier

A two-compartment fractional derivative model for Propofol diffusion in anesthesia

This paper presents the initial steps towards the development of a compartmental model for Propofol diffusion in the human body using concepts from fractional calculus. The model presented here preserves the mass balance, therefore it maintains the link between physiological and mathematical parameters. The final purpose of this model is to predict drug pharmacokinetics and pharmacodynamics during general anesthesia. A comparison of the proposed a fractional order model and the integer order compartmental model of Propofol shows good agreement between them.

A Three-Year Feedback Study of a Remote Laboratory Used in Control Engineering Studies

This paper discusses the results of a feedback study for a remote laboratory used in the education of control engineering students. The goal is to show the effectiveness of the remote laboratory on examination results. To provide an overview, the two applications of the remote laboratory are addressed: 1) the Stewart platform, and 2) the quadruple water tank system. Combining both applications allows a broad spectrum of practical examples featuring challenging control aspects such as multiple-input–multiple-output control, decoupling, non-minimum phase systems, open-loop unstable systems, and PID control design. The remote laboratory feedback study was performed using a five-point Likert-type scale survey to elicit the students’ level of satisfaction with the laboratory. Three years of student examination results were also studied to compare performance before and after integrating the remote laboratory. In the first of these years there was no use of the remote laboratory. In the second year, the remote laboratory was introduced on a voluntary basis, and in the third year the remote laboratory was obligatory. Student feedback indicates that the remote laboratory needs further development to counter its limitations. A major conclusion of the survey was that there is keen interest in the remote laboratory to provide practical experience in the training of a control engineer. It can be concluded that the remote laboratory has a positive effect on student examination results.

Nonlinear Predictive Control of Shank Movement Around the Knee Joint

Abstract This paper derives an elementary mathematical model for the movement of the shank around the knee joint. The resulting equation of motion is a nonlinear differential equation. The control purpose is to regulate the angle of the shank by manipulating the torque applied to the knee joint. This model has applications in dynamical knee rigs and active orthoses. As the system is nonlinear with a rather larger range of motion, a nonlinear predictive controller is designed. The implemented control strategy is NEPSAC (Nonlinear Extended Prediction Self-Adaptive Control). The effect of the filter parameter and the prediction horizon in the NEPSAC algorithm are investigated using MATLAB/SIMULINK. The results show that for both the filter parameter and the prediction horizon an optimal value can be chosen as a trade-off between robustness and fastness of the response.

Emerging Tools for Quantifying Unconscious Analgesia: Fractional-Order Impedance Models

This paper presents the application of model-based predictive control (MPC) in combination with a sensor for the measurement of analgesia (pain relief) in an unconscious patient in order to control the level of anesthesia. The MPC strategy uses fractional-order impedance models (FOIMs) to model the diffusion process that occurs in the human body when an analgesic drug is taken up. Based on this control strategy an early dawn concept of the pain sensor is developed. The grand challenges that coincide with this development include identification of the patient model, validation of the pain sensor, and validation of the effect of the analgesic drug.

Remote laboratory: A novel tool for control engineering laboratories

This paper explores how two scientific experiments in control engineering can be performed remotely over the Internet: (1) Ball and Plate system; and (2) Water Tank system. The main objective is offering students insights about identification of real processes and auto-tuning techniques. At Ghent University real-time remote access laboratories, which allow students to implement, compile and run controllers remotely from home, have been developed. Students can watch the experiment in real time and effectively use the laboratories full functionality through Virtual Private Network (VPN) connection across the Internet.

Formation control strategies for emulation of field covering

This work aims at investigating the subject of formation control, i.e. controlling a group of mobile robots to make them form a specific geometric pattern and, additionally, to have them maintain this pattern while moving. In view of this two-fold scope, we have conducted experiments using mobile robots equipped with an on-board camera that, by means of a stepper motor, can rotate over 360°. Within this setting, the rotatable camera functions as a decentralized sensor that gives the position and orientation of the follower with respect to the leader. Given the lack of communication between the robots, we use a combination of PID-PI cascade control and a PI feedback loop to have them form and sustain a triangular leader-follower formation. As a way of illustrating the concept, we will discuss a practical application whereby a formation of combine harvester machines is used to harvest crops.

Decoupled PID control with gain adaptation for a cycling dynamic knee rig

This paper presents the deployment and validation of a control structure for a dynamic knee rig used to gain insight into the kinematics and kinetics of the human knee joint. The dynamic knee rig is able to implement a bicycle motion onto the knee joint using five different actuators. Two actuators control the position of the ankle, one actuator mimics the quadriceps force and two more actuators impose the hamstring forces. Identification of the subsystems of the rig is performed using the ARX method. The obtained mathematical models are used to tune the controller parameters of the proportional-integral-derivative (PID) controllers which control each subsystem. Apart from PID controllers, the control structure also contains feedforward action, decouplers and gain adaptation mechanism. The controllers are tested using a mechanical hinge as knee joint The results indicate a good trajectory tracking performance of the controllers.

A Multi Agent System for Precision Agriculture

This chapter investigates the use of a multi agent system for precision agriculture. Precision agriculture refers to the management of farm operations based on observation, measurement and fast response to inter- and intra-field variability in crops. Important characteristics in this application are path following, disturbance rejection and obstacle avoidance of which path following is addressed in this chapter. This study combines the degree of freedom of aerial vehicles with the location precision of ground vehicles in order to achieve a common goal. The multi agent system in this study consists of a quadrotor as the aerial agent and tracked robots with a rotary camera as the agents on the ground to achieve the common task of following a predefined path while maintaining the formation. This research uses a combination of low-level PID cascade control for the ground vehicles with a high level predictive control for the aerial agent to ensure optimal control of the positions of the ground robots and the quadrotors. A series of proof-of-concept experiments for this novel combined control strategy are performed. Simulation and experimental results suggest that the proposed control system is able to maintain the formation of the ground vehicles and provide a good tracking of the ground vehicles by the quadrotor. Because of the setup of the described system, it has also potential applications in traffic control and analysis, search and rescue operations, etc.

Fractional order impedance models as rising tools for quantification of unconscious analgesia

This research focuses on modeling the diffusion process that occurs in the human body when an analgesic drug is taken up, by using fractional-order impedance models (FOIMs). We discuss the measurement of a suitable feedback signal that can be used in a model-based control strategy. With this knowledge an early dawn concept of a pain sensor is presented. The major challenges that are encountered during this development consist of identification of the patient model, validation of the pain sensor and validation of the effect of the analgesic drug.

Fractional-order PID design: Towards transition from state-of-art to state-of-use

This paper presents a new tuning method for fractional-order (FO)PID controllers to simplify current tuning and make FOPID controllers more convenient for industry, i.e. facilitate transition from state-of-art to state-of-use. The number of tuning parameters is reduced from five to three based on popular specification settings for PID controllers without the need for reduced process models which introduce modeling errors. A test batch of 133 simulated processes and two real-life processes are used to test the presented method. A comparative study between the new method and the established CRONE controller, quantifies the performance. The conclusion states that the new method gives fractional controllers with similar performances as the current methods but with a significantly decreased tuning complexity making FOPID controllers more acceptable to industry.