Design and Real-time Implementation of Robust Control Schemes for Twin Rotor Multivariable System

Abstract

In the real world, many physical systems are not only nonlinear but also highly uncertain. Hence, it is necessary to design robust controllers for these systems to achieve efficient set-point tracking and disturbance rejection. A lot of research works have been directed in the past to develop controllers for Multi-Input Multi-Output (MIMO) systems. MIMO systems are more complex than a Single-Input Single-Output (SISO) systems, because in the former, loop interactions exist among different control loops. Thus, opportunities lie in developing control strategies for uncertain non-linear MIMO systems like magnetic levitation system, helicopter system, under water vehicle, and twin rotor system etc. Among the non-linear MIMO systems, helicopter system is considered as a one of the most popular complex MIMO system. As the model of the helicopter system is so expensive, various researchers have chosen the Twin Rotor MIMO System (TRMS) as it is a laboratory prototype model of helicopter system. Basically, aerodynamic control of helicopter and TRMS are little bit different. Aerodynamic control of helicopter varies the angle of the rotor blades whereas TRMS varies the speed of the motors. However, because of the external disturbances, model uncertainties and inherent cross-coupling effects, designing a suitable controller for TRMS becomes a challenging task. As the dynamics of the TRMS is complex due to its nonlinearities, cross coupling; System Identification (SI) method is adopted for modelling of the system. SI toolbox constructs mathematical models of dynamic systems from measured input-output data. SI toolbox of MATLAB is a good way of obtaining models for systems that are difficult to model. Autoregressive Moving Average Model with Exogenous inputs (ARMAX) model is very often used in performing system identification. After obtaining suitable model of the TRMS by employing SI, the thesis explores design of efficient robust control schemes. Several robust control approaches are exploited to realise these controllers as described below. The control of TRMS is difficult to operate because of external disturbances and uncertainties. Moreover, there is a coupling between the pitch and yaw channels, making it more difficult for the two channels to be controlled independently. For the control of pitch and yaw channels for TRMS, an Internal Model Control (IMC) based Proportional Integral Derivative (PID) controller is proposed in this thesis. In order to analyze the performance, robust PID controller based on Linear Quadratic Regulator-Linear Matrix Inequality (LQR-LMI) approach is compared with the proposed controller. From the Obtained results, it is confirme that IMC based PID controller exhibits superior performance than the L QR LMI based PID controller, in terms of improved attitude angle tracking and disturbance rejection ability. Considering the difficulties encountered in TRMS, a H¥ controller with two Degree of Freedom (TDOF) is designed to handle the mode lun certain ties and externaldi sturbances. For handling both of these, TRM Suses mixed sensitivity approach. For performance analysis, the proposed H¥ controller is compared with L QR-LMI based robust PID controller and IMC based PID controller in MATLAB/Simulink and then in real platform. Based on the results, the proposed controller can be concluded to be more robust, faster in tracking performance, and more accurate in disturbance attenuation than LQR-LMI-based robust PID controller sand IMC based PID controller. A TDOF-IMC based Generalized Active Disturbance Rejection Controller (GADRC) for TRMS is proposed in this thesis. Similar to ADRC, GADR Cuses the Extended State Observer (ESO) to estimate in ternalun certainties and lumped disturbances as external disturbances. GADRC is an extremely advantageous controller synthesis approach that can utilize any available data concerning the system. Further more, GADRC can also beutilized for non-minimum phase a swellasun stable processes that are extremely challenging to control using the ADRC. The GDRC structure is converted in to the renowned TDOF-IMC structure for ease of analysis and tuning. Further more, it is shown that the tuning b and widths for the set point filter and the disturbance rejection filter in TDOF-IM Care the inverse of the two time constants, thus tuning be comes easieror those practitioners familiar with the IMC method. Among all the robust controller proposed in the thesis, GADRC controller is found to be the most effective one.