Anindita Sengupta

Nonparametric modeling of glucose-insulin process in IDDM patient using Hammerstein-Wiener model

This paper deals with an identification problem of modeling a nonlinear dynamic system of multivariable glucose-insulin process in an IDDM patient. Out of many model structures that can represent a nonlinear process effectively; the Hammerstein-Wiener model has attracted a lot of attention. The present work proposes a generalized identification method of Hammer stein-Wiener model from the input-output data of multivariable nonlinear glucose-insulin process. The present algorithm consists of a three-block (LNL) realization. For the multivariable system, the first and third blocks are standard impulse response filter (TRF) realization applied to an equivalent linear system using adaptive recursive least square (ARLS) algorithms. In the second block, i.e. the nonlinear part, ARLS algorithms have been used to solve up to second order kernels of Volterra equations with extended input vector consisting of cross components as well. The input-output data taken from the simulated nonlinear process have been used to identify the system with a filter memory length of M=3 and the validation results have shown good fit and in concordance with predicted output.

Servomechanism for periodic reference input: Discrete wavelet transform-based repetitive controller

The presented work explored the possibility of developing a discrete wavelet transform-based repetitive control (DWTRC) strategy along with a proportional controller for a liquid level system (LLS) to track a periodic reference signal. The modified DWTRC with a P control system is a type of servo mechanism, effective for periodic reference input. The entire effort derives from two distinct parts. First, the tracking performance and stability measure of the system with inclusion of a repetitive controller (RC) in presence of P controller that has been studied and compared with conventional P and PI controllers in both time and frequency domains, subjected to periodic command to establish the usefulness of the former. The most obvious disadvantage of the RC with P control is the requirement of large memory for the presence of RC, and the memory requirement linearly increases with both repetition time and sampling frequency. Second, to overcome this difficulty, deployment of the discrete wavelet transform (DWT) in the RC loop is done to reduce the problem of memory requirement, and ensures efficient and high-speed computation due to the filtering and data reduction property of DWT. Finally, the effectiveness of DWTRC is validated through saving memory and optimizing the tracking error with variation of the wavelet order and its decomposition level.

Wavelet based noise reduction of Liquid level system using minimum description length criterion

The purpose of this paper is to employ wavelet based noise removal technique for Liquid level system (LLS) to remove the measurement noise from differential pressure transmitter (DPT) output indicating the level of a process tank via discrete wavelet transform (DWT). The choice of which wavelet to use for denoising is of critical importance because the wavelet affects reconstructed signal quality and the design of the system as a whole. The minimum description length criterion permits to select not only the suitable wavelet filter but also the best number of wavelet retained coefficients for signal reconstruction. The hard thresholding methods typically retain a very small number of coefficients, and the results are often smoothed. By retaining a slightly larger amount of coefficients and shrinking them, the soft thresholding methods usually give better results than the hard thresholding methods to de-noise any signal. The available response is compared with that of conventional Butterworth filtering method.

Application of MRPID Controller on Liquid Level System: A Performance study

Comparative study of the performance of Wavelet based Multi Resolution PID (MRPID) controller over conventional PID controller for a Liquid level system is presented in this paper. In the proposed wavelet based PID controller, the discrete wavelet transform (DWT) is used to decompose the error signal available from the set point signal and measured signal into different frequency components at various levels. Then the calculated DWT coefficients are multiplied with their respective gains to generate an overall control signal. Both the controllers are simulated in LabVIEW software. After implementation of software model successfully, plant is realized in hardware with controller simulated in PC. The performances of the wavelet-based MRPID controller are found to be superior to that of the conventional PID controller even in the presence of the white noise.

Selection of best wavelet for discrete wavelet transform based PID controller connected with liquid level system and its performances analysis

In this paper, the performance of discrete wavelet transform (DWT) based PID controller is compared to the conventional proportional-integral-derivative (PID) controller applied to Liquid level system (LLS). Here, DWT is used to decompose the actuating signal into different scales. The high scale gives the approximate or low frequency information and low scale gives the detail or high frequency information contained in the actuating signal. Based on this frequency information, the decomposed coefficients of each level are scaled by their respective gains and then added together to generate the final control signal. The choice of wavelet to decompose the error signal is an important factor because the wavelet affects the reconstructed signal quality and the design of the system as a whole. The minimum description length criterion permits to select not only the suitable wavelet filter but also the optimum number of wavelet coefficients retained for the signal reconstruction. Simulation results show that the proposed controller has faster transient response and more satisfactory operation under steady-state conditions with respect to PID controller. Simulation is done in National Instruments LabVIEW software.

Estimation of tire-road friction coefficient and frictional force for Active Vehicle safety system

An increasing number of accidents of Vehicles has led to the study and design of Active safety systems in modern automobiles. For the purpose, a number of sensors for the measurement of Vehicle yaw, wheel velocities and acceleration are deployed. However, some key parameters like slip angle and frictional forces are hard to be determined using sensors and also cost prohibitive. Estimation of friction coefficient and frictional forces has been of wide importance for the design of Active safety systems as the information is required for the design of efficient control system. Besides, the system being highly nonlinear in nature, linearized estimation technique may lead to high approximation errors. This paper presents an unscented Kalman Filter based estimation algorithm for a specific nonlinear tire model for the estimation of friction coefficient and lateral and longitudinal frictional forces.

Yarn hairiness evaluation using image processing

Yarn hairiness or amounts of released fibers from the staple yarn are one of the main aspects to detect the yarn quality. This paper illustrates the method of finding yarn hairiness using image processing techniques with a low-cost USB Web Camera in association with a yarn moving arrangement. The steps involved with the image processing comprises of segmenting the yarn core from the image and fibre extraction. The yarn hairiness found by the proposed method are quantised by different indices, such as yarn hair area index and yarn hair length index and count of hair tips at specific distances from the central line of the yarn core. The future goal of this work is to acquire the image of the yarn from a spinning frame with the help of a high resolution web camera and process the same with advanced software to characterize the yarn which in turn dictates the fabric appearance in terms of hairiness.

Filter design for tracking of ballistic target missile using seeker measurements with time lag

Here a method has been proposed to estimate the realtime state using the sensor measurement which contains lag due the mechanism of data collection and transmission to the estimator for realtime state estimation. The error due to the measurement lag can be handle by incorporating the sensor dynamics in the system model. In this paper A extended Kalman filter has been designed to estimate relative kinematic of a ballistic target missile w.r.t. interceptor missile using noisy and lagged seeker measurements in exo atmospheric region. The seeker dynamics has been modeled and included in the system model of the estimator design to handle lag present in the seeker measurement due to processing and transmitting data which are obtained from different reference frame.

Parameter selection strategy for robust sliding mode controller and its implementation in real time system

This paper presents the robust controller design strategy for nonlinear systems. A novel controller parameter selection methodology is incorporated in the design procedure so that the system states can quickly reach the defined desired level. The experiment is first performed in a simulation environment and the parametric values obtained from the experiment are applied in a real time physical system. The dynamic performance of the system with the variation of parameter values are also incorporated in the paper.

Finite dimensional robust repetitive controller for tracking periodic reference input

The presented work demonstrates in steps the way of development and deployment of Finite Dimensional Robust Repetitive Controller (FDRRC) starting from the conventional Repetitive Controller (RC) which is actually an infinite dimensional system. By the virtue of being infinite dimensional RC can track any periodic reference input but many times degrading stability due to the small period mismatch between controller period and actual period of disturbance and also in high frequency applications. To minimize the problem of instability in high frequency, a low pass filter is introduced in the RC loop which in turn limits the infinite dimensional RC in to Finite Dimensional RC (FDRC) and to deal with small period uncertainties, an attempt is taken to explore the design and implementation of multiple loops Robust Repetitive Controller (RRC). Combination of the said strategies results in FDRRC. The new idea of FDRRC is utilized and experienced through the a 1st order system in this paper and every time a comparison is carried out to show the necessity of development in the considered controller models. Finally the paper tries to emphasise over the effectiveness of the FDRRC based on results obtained.