Parameter estimation is a first and foremost task to design a proper mathematical model. Outliers in a data can often lead to improper parameter estimation. In present work a novel technique using interval constraint satisfaction technique is suggested for getting robust parameter estimation of a system using linear orthogonal regression. In this proposed method we used M estimator with tukey’s biweight function and modified it for orthogonal regression. Tukey’s biweight is generally computed using IRLS which has computational issues of convergence and local minima. Here, benefits of interval analysis based method is utilized to develop the proposed method. Two well-known examples of datasets are used to validate the proposed method.

Quantitative feedback theory (QFT) is a frequency domain based robust control design techniques [1]. In this paper, the piecewise linear approximation based linear programming (PLA-LP) algorithms presented in [6] are extended for automatic synthesis of robust PID controller in the QFT framework. Firstly, the λ-formulations for the parallel or interacting form of the PID controller are derived. Then, the λ-formulations are incorporated into the QFT synthesis algorithm proposed in [6], and solved to global optimality using a mixed-integer programming (MIP) branch and bound solver. The proposed robust QFT-PID synthesis algorithm is demonstrated on two benchmark examples. Compared to existing methods, the PLA-LP based QFT-PID algorithm is faster by a couple of orders of magnitude, and also gives considerable reductions in the controller gain.

This paper experimentally assesses the performance of recent QFT robust control synthesis algorithms (Makwana and Nataraj in Proceedings of first IEEE international conference on industrial instrumentation and control (ICIC 2015), Pune, India, 2015), called PLA-LP algorithms, on electromechanical systems. An industrial plant emulator that is representative of modern electromechanical plants is used for the experimental studies. The industrial plant emulator offers many control challenges in the form of plant uncertainty, output disturbances, backlash nonlinearity, and friction. To perform in face of these challenges, a robust control system is designed using the PLA-LP algorithms. The performance assurance capability of the control system is then investigated via real-time experiments on the industrial plant emulator. The emulator is experimentally subjected to a variety of tests, such as plant uncertainty, setpoint changes, input and output disturbances, and their various combinations. The obtained experimental results show that the PLA-LP design algorithms indeed assure that the control system performs as per the given robust control specifications.

Prefilter synthesis is one of the important design steps of Horowitz's quantitative feedback theory (QFT) to robust feedback system synthesis. The prefilter is designed to achieve tracking specifications. In this paper, a new, computationally efficient approach for automation of the prefilter design step is proposed. A linear programming based algorithm is proposed for the automatic synthesis of fixed structure QFT prefilter. In present work, prefilter synthesis problem is posed as a piecewise linear approximation based linear programming problem and solved with GUROBI optimizer solver. The proposed method is validated using a benchmark problem and a simple, low order prefilter is obtained in quick time.

Robust controller synthesis is of great practical interest and its automation is a key concern in control system design. Automatic controller synthesis is still a open problem. Presently available automated synthesis methods of QFT controller design suffer from convergence to local minima, requirement of initial seed controller, limitation of maximum controller parameters etc,. In this paper a new, fast and efficient method has been proposed for automated synthesis of a fixed structure quantitative feedback theory (QFT) controller. The controller synthesis problem is posed as a piecewise linear approximation based linear programming problem and solved with GUROBI optimizer solver. The proposed algorithms converges to approximate global optimum and supports synthesis of controllers having large number of parameters in very less time. The proposed method is tested on two benchmark problems, and simple, low order controllers are successfully obtained in quick time.

Parameter estimation is crucial for model identification in control engineering and in other disciplines. In present work a new method based on interval constraint satisfaction technique is proposed for obtaining parameter estimation of a system using linear orthogonal regression. In orthogonal regression, errors in both the measured variable and response variable are taken into account. The power of interval analysis based method is explored to develop proposed method. Two well known examples of datasets are used to validate the proposed method.