Michail Petrov

Fuzzy PID control of nonlinear plants

It is difficult to achieve efficient control of time variable and nonlinear plants with conventional PID controllers. A method of designing a nonlinear fuzzy PID controller is presented. The nonlinear fuzzy PID controller could be applied successfully in control systems with various nonlinearities. The fuzzy PID controller can be viewed as a natural similarity to the conventional PID controller. This paper describes the structure and the design aspects of a fuzzy PID controller based on Sugeonos fuzzy technique with fuzzy-neural implementation. There are two possibilities to obtain a three-term fuzzy PID controller similar to the conventional digital PID controller. The first one is a velocity type fuzzy PID controller and the second one is the positioning type fuzzy PID controller. The antecedent part of the applied fuzzy rules contains a linear function, similar to the discrete equation of the corresponding conventional PID controller. The paper contains the structure and description of this implementation as well as investigations toward the applied fuzzy PID control algorithms. The simulations demonstrate satisfactory results of these performances and implementations applied to a nonlinear plant composed by two cascaded water tanks with level control.

Intelligent Sensor-Based Navigation and Control of Mobile Robot in a Partially Known Environment

Abstract Navigation and collision avoidance are major areas of research in mobile robotics that involve varying degrees of uncertainty. In this paper, a new approach is proposed for navigation and control of a wheeled mobile robot in a partially known environment. A collision-free path is calculated using an efficient neural motion planner. An additional fuzzy logic based navigation strategy is developed to adjust the moving information in the cases when the mobile robot encounters unexpected obstacles. The output of the navigation level is transformed into a time indexed data sequence which is fed into a tracking controller that takes into account the complete dynamics of the mobile base. The locomotion control structure is based on the integration of a kinematic controller and an adaptive fuzzy-net torque controller.

Intelligent Motion Planning and Control for a Simulated Mobile Robot

Abstract A complete motion planning and control procedure for mobile robot is presented. A collision-free path is calculated using a neural-net motion planner. The output of the planner is then transformed into a time indexed data sequence which is fed into a tracking controller that takes into account the complete dynamics of the mobile base. A locomotion control structure based on the integration of a kinematic controller and an adaptive fuzzy-net torque controller is proposed. An evolutionary feedback-error-learning method for automatic elicitation of knowledge in the form of fuzzy if-then rules is developed. The results of the simulations show the effectiveness of the proposed approach.

Implementations of a Hammerstein fuzzy-neural model for predictive control of a lyophilization plant

This paper describes two methodologies for implementation of Hammerstein model by using different input-output representations into model predictive control schemes. The model nonlinearity is easily approximated using a simple Takagi-Sugeno inference, while the linear parts are flexibly introduced. As optimization procedures for predictive control are used a standard gradient optimization method and an implementation of Hildreth Quadratic Programming. A comparison between the proposed control strategies is made by simulation experiments for control of nonlinear lyophilization plant.

MODEL PREDICTIVE CONTROL OF A LYOPHILIZATION PLANT: A SIMPLIFIED APPROACH USING WIENER AND HAMMERSTEIN SYSTEMS

Lyophilization process is widely used in pharmaceutical industries, preparing stable dried medications and important biopreparations, so they remain stable and easier to store at room temperature. Since a lyophilization cycle involves high energy demands, an improved control strategy has to be used in order to minimize the operating costs. This paper deals with the design methodology of nonlinear model predictive controllers for lyophilization plant. The controllers are based on fuzzy-neural predictive models and simplified gradient optimization algorithm. As predictive models, fuzzy-neural implementations of Hammerstein and Wiener-Hammerstein systems are used. Such structures provide fast and reliable system identification using small number of parameters which reduces the computational burden during the optimization procedure. The potential benefits of the proposed approaches are demonstrated by simulation experiments.

Implicit GPC Based on Semi Fuzzy Neural Network Model

The model in Model Predictive Control (MPC) takes the central place. Therefore, it is very important to find a predictive model that effectively describes the behavior of the system and can easily be incorporated into MPC algorithm. In this paper it is presented implicit Generalized Predictive Controller (GPC) based on Semi Fuzzy Neural Network (SFNN) model. This kind of model works with reduced number of the fuzzy rules and respectively has low computational burden, which make it suitable for real-time applications like predictive controllers. Firstly, to demonstrate the potentials of the SFNN model test experiments with two benchmark chaotic systems - Mackey-Glass and Rossler chaotic time series are studied. After that, the SFNN model is incorporated in GPC and its efficiency is tested by simulation experiments in MATLAB environment to control a Continuous Stirred Tank Reactor (CSTR).

Fuzzy-neural model predictive control of a building heating system

Abstract This paper describes the development of a Model Predictive Controller with supervision control of a building heating system. A fuzzy–neural model and optimizing procedure as a part of a nonlinear predictive controller are utilized on-line to determine the future values of control actions based on dependence between outdoor and indoor temperatures. A learning algorithm for parameters in fuzzy-neural implementation of the predictive model is additionally applied. Simulation results with a model of a single room heating system demonstrate that a better system performance can be achieved in comparison to classical PID control.

Adaptive supervisory tuning of nonlinear model predictive controller for a heat exchanger

Abstract It is presented in this paper an adaptive predictive supervisory algorithm to the temperature control of a heating system with a heat exchanger. The nonlinear predictive control strategy is designed on the basis of a Takagi-Sugeno fuzzy-neural model and a simple optimization procedure. An additional supervisory level in the control system is introduced for adaptive tuning of a weighting factor in the predefined optimization criterion. Using the proposed algorithm a higher system performance can be achieved which leads to reduction of the energy consumption into the heating system. The proposed approach is studied by experimental simulations to control a temperature in the heating system.

Fuzzy-neural predictive control using Levenberg-Marquardt optimization approach

It is proposed in this paper a study on the influence of the Levenberg-Marquardt optimization approach for computation of the control actions in Nonlinear Model Predictive Controller. To predict the future plant behavior, a classical Takagi-Sugeno inference is used. A comparison by applying the Gradient descent and the Newton-Raphson optimization approaches is made. The efficiency of the proposed optimization strategies is demonstrated by experiments in MATLAB environment to control a Continuous Stirred Tank Reactor.

Parallel distributed neuro-fuzzy model predictive controller applied to a hydro turbine generator

Model predictive control (MPC) has been considered as the most important development in the area of process control in the last two decades. This paper addresses the issue of controlling a nonlinear plant by the use of the nonlinear model predictive control formulation. To handle the nonlinearities, a Takagi-Sugeno neuro-fuzzy model is suggested as a means to model the plant with nonlinearities depending on the operating region. The developed model is used as a predictive model for a parallel distributed model predictive control algorithm. In this paper, the parallel distributed neuro-fuzzy model predictive controller has been proposed to control a non-linear control system of a hydro turbine generator. The proposed technique has been tested and evaluated using this simulated industrial plant.