Siti Fauziah Toha

Dynamic Nonlinear Inverse-Model Based Control of a Twin Rotor System Using Adaptive Neuro-fuzzy Inference System

A dynamic control system design has been a great demand in the control engineering community, with many applications particularly in the field of flight control. This paper presents investigations into the development of a dynamic nonlinear inverse-model based control of a twin rotor multi-input multi-output system (TRMS). The TRMS is an aerodynamic test rig representing the control challenges of modern air vehicle. A model inversion control with the developed adaptive model is applied to the system. An adaptive neuro-fuzzy inference system (ANFIS) is augmented with the control system to improve the control response. To demonstrate the applicability of the methods, a simulated hovering motion of the TRMS, derived from experimental data is considered in order to evaluate the tracking properties and robustness capacities of the inverse- model control technique.

PID-Based Control of a Single-Link Flexible Manipulator in Vertical Motion with Genetic Optimisation

This paper presents an investigation into dynamic simulation and controller optimization based on genetic algorithms (GAs) for a single-link flexible manipulator system in vertical plane motion. The dynamic model of the system is derived using the Lagrange equation and discretised using the finite difference (FD) method. GA optimization is used to optimize the parameters of the proportional-integral-derivative (PID) based controllers for control of rigid-body and flexible motion dynamics of the system. The important point is to evaluate the range of PID parameter which used in the GAs programmed to find the best value of this parameter. Comparative performance assessment of the control approaches are presented and discussed in the time and the frequency domains.

Real-coded genetic algorithm for parametric modelling of a TRMS

This paper present a novel and scrutinized parametric modeling of a laboratory scale helicopter, a twin rotor multi input multi output system (TRMS), by employing a real-coded genetic algorithm (GA) technique. The main goal of this work is to emphasise the potential benefits of this architecture for real system identification. Instead of working on the conventional bit by bit operation, both the crossover and mutation operators are real-valued. The effectiveness of the proposed algorithm is demonstrated in comparison to a binary-coded GA in modelling the TRMS. A complete system identification procedure has been carried out, from experimental design to model validation using a laboratory-scale helicopter. In this case, the identified model is characterized by a fourth order linear ARMA structure which describes with very high precision the hovering motion of a TRMS. The TRMS can be perceived as a static test rig for an air vehicle with formidable control challenges. Therefore, an analysis of modeling of nonlinear aerodynamic function is needed and carried out in both time and frequency domains based on observed input and output data. Experimental results are obtained using a laboratory set-up system, confirming the viability and effectiveness of the proposed methodology.

MLP and Elman recurrent neural network modelling for the TRMS

This paper presents a scrutinized investigation on system identification using artificial neural network (ANNs). The main goal for this work is to emphasis the potential benefits of this architecture for real system identification. Among the most prevalent networks are multi-layered perceptron NNs using Levenberg-Marquardt (LM) training algorithm and Elman recurrent NNs. These methods are used for the identification of a twin rotor multi-input multi-output system (TRMS). The TRMS can be perceived as a static test rig for an air vehicle with formidable control challenges. Therefore, an analysis in modeling of nonlinear aerodynamic function is needed and carried out in both time and frequency domains based on observed input and output data. Experimental results are obtained using a laboratory set-up system, confirming the viability and effectiveness of the proposed methodology.

ANFIS modelling of a twin rotor system using particle swarm optimisation and RLS

Artificial intelligence techniques, such as neural networks and fuzzy logic have shown promising results for modelling of nonlinear systems whilst traditional approaches are rather insufficient due to difficulty in modelling of highly nonlinear components in the system. A laboratory set-up that resembles the behaviour of a helicopter, namely twin rotor multi-input multi-output system (TRMS) is used as an experimental rig in this research. An adaptive neuro-fuzzy inference system (ANFIS) tuned by particle swarm optimization (PSO) algorithm is developed in search for non-parametric model for the TRMS. The antecedent parameters of the ANFIS are optimized by a PSO algorithm and the consequent parameters are updated using recursive least squares (RLS). The results show that the proposed technique has better convergence and better performance in modeling of a nonlinear process. The identified model is justified and validated in both time domain and frequency domain

Genetic Algorithm Optimisation for Fuzzy Control of Wheelchair Lifting and Balancing

In this paper, an optimisation technique is adopted to manipulate the input and output scaling of a fuzzy logic controller for lifting the front wheels of a wheelchair and stabilizing the wheelchair in two-wheeled mode. A virtual wheelchair (WC) model is developed within Visual Nastran (VN) software environment where the model is further linked with Matlab/Simulink for control purposes. The lifting of the chair is done by transforming the first link, attached to the front wheels (casters) to the upright position while maintaining stability of the second link where the payload is attached. General rules of thumb allow heuristic tuning of the parameters but a proper optimisation mechanism will perform better. Genetic Algorithm is used to control the two-wheeled wheelchair and results show that the optimised parameters give better system performance.

A Hybrid Control Scheme for a Twin Rotor System with Multi Objective Genetic Algorithm

Control of vibration suppressions is crucial for applications in engineering particularly in the area of aircraft system. A hybrid control approach encompasses of using a feedforward intelligent command shaping technique with combination of intelligent PID feedback control is presented in this paper. The advantage of using command shaping is to reduce system vibration. However, it can cause delay in system response resulting to a conflict between vibration suppression and rise time response. Multi objective genetic algorithm is employ in this work to determine a set of solutions for the amplitudes and corresponding time locations of impulses on an extra sensitive (EI) command shaping as well as gain parameters for the PID controller. The effectiveness of the proposed technique is assessed both in the time domain and the frequency domain. Moreover, a comparative assessment of the performance of the technique with the system response and unshaped finite step input is presented.

Augmented feedforward and feedback control of a twin rotor system using real-coded MOGA

Vibration suppression is crucial for applications in engineering particularly for aircraft systems. A hybrid control approach comprising a feedforward intelligent command shaping technique inverse-model based PID feedback control is presented in this paper. The proposed augmented control scheme is used to control both the flexible motion and rigid body dynamics of a twin rotor multi-input multi-optput system (TRMS). The advantage of using command shaping is to reduce system vibration. However, it can cause delay in system. Furthermore, performance requirements based on tracking error, rise time, settling time, percentage overshoot and steady state error are often found to be conflicting with one another in most flexible systems. Therefore real-coded multi objective genetic algorithm is employed in this work to compromise the problems and determine a set of solutions for the amplitudes and corresponding time locations of impulses on an extra sensitive (EI), four-impulse sequence command shaper as well as gain parameters for the PID controller. The effectiveness of the proposed technique is assessed both in the time domain and the frequency domain. Moreover, a comparative assessment of the performance of the technique with the system response and unshaped finite step input is presented.

Parametric Modelling of a TRMS Using Dynamic Spread Factor Particle Swarm Optimisation

System identification in vibrating environments has been a matter of concern for researchers in many disciplines of science and engineering. In this paper, a sound approach for a Twin Rotor Multi-input Multi-Output System (TRMS) parametric modeling is proposed based on dynamic spread factor particle swarm optimization. Particle swarm optimization (PSO) is demonstrated as an efficient global search method for nonlinear complex systems without any a priory knowledge of the system structure. The proposed method formulates a modified inertia weight algorithm by using a dynamic spread factor (SF). The inertia weight plays an important role in terms of balancing both the global and local search. Thus, the usage of dynamic SF is proved experimentally to satisfy main issues of using basic PSO that are trapped in local optima and preservation of diversity. Results in both time and frequency domains portray a very good parametric model that mimic well the behavior of a TRMS. Validation tests clearly show the effectiveness of the algorithm considered in this work.

Inverse model based control for a twin rotor system

The use of active control technique has intensified in various control applications, particularly in the field of aircraft systems. A laboratory set-up system which resembles the behaviour of a helicopter, namely twin rotor multi-input multi-output system (TRMS) is used as an experimental rig in this research. This paper presents an investigation using inverse model control for the TRMS. The control techniques embraced in this work are direct inverse-model control, augmented PID with feedforward inverse-model control and augmented PID with feedback inverse-model control. Particle swarm optimization (PSO) method is used to tune the parameter of PID controller. To demonstrate the applicability of the methods, a simulated hovering motion of the TRMS, derived from experimental data is considered. The proposed inverse model based controller is shown to be capable of handling both systems dynamic as well as rigid body motion of the system, providing good overall system performance.