Raja Kamil

Automation of anaesthesia: a review on multivariable control

Anaesthesia is a multivariable problem where a combination of drugs are used to induce desired hypnotic, analgesia and immobility states. The automation of anaesthesia may improve the safety and cost-effectiveness of anaesthesia. However, the realization of a safe and reliable multivariable closed-loop control of anaesthesia is yet to be achieved due to a manifold of challenges. In this paper, several significant challenges in automation of anaesthesia are discussed, namely model uncertainty, controlled variables, closed-loop application and dependability. The increasingly reliable measurement device, robust and adaptive controller, and better fault tolerance strategy are paving the way for automation of anaesthesia.

Comparison of Performance and Computational Complexity of Nonlinear Active Noise Control Algorithms

Research on nonlinear active noise control (NANC) revolves around the investigation of the sources of nonlinearity as well as the performance and computational load of the nonlinear algorithms. The nonlinear sources could originate from the noise process, primary and secondary propagation paths, and actuators consisting of loudspeaker, microphone or amplifier. Several NANCs including Volterra filtered-x least mean square (VFXLMS), bilinear filtered-x least mean square (BFXLMS), and filtered-s least mean square (FSLMS) have been utilized to overcome these nonlinearities effects. However, the relative performance and computational complexities of these algorithm in comparison to FXLMS algorithm have not been carefully studied. In this paper, systematic comparisons of the FXLMS against the nonlinear algorithms are evaluated in overcoming various nonlinearity sources. The evaluation of the algorithms performance is standardized in terms of the normalized mean square error while the computational complexity is calculated based on the number of multiplications and additions in a single iteration. Computer simulations show that the performance of the FXLMS is more than 80% of the most effective nonlinear algorithm for each type of nonlinearity sources at the fraction of computational load. The simulation results also suggest that it is more advantageous to use FXLMS for practical implementation of NANC.

Loudspeaker nonlinearity compensation with inverse tangent hyperbolic function-based predistorter for active noise control

In active noise control (ANC), the performance of the filtered-x least mean squares (FXLMS) algorithm is degraded by the saturation of the loudspeaker in the secondary path. Predistortion is a linearization technique commonly used in signal processing applications to compensate for saturation nonlinearity. The design of the predistorter (PD) requires the use of direct measurement from the output of the nonlinear element. However, in ANC applications, direct measurement from the loudspeaker output is not available. Therefore, a conventional PD design approach cannot be directly applied. In this paper, a new PD-based compensation technique based on the inverse model of the loudspeaker nonlinearity is proposed. The PD is represented by an approximated memory-less inverse tangent hyperbolic function (ITHF). The approximated ITHF is scaled by a pre-identified parameter, which represents the loudspeaker nonlinearity strength. This parameter can be obtained by modelling the secondary path using a proposed block-oriented Hammerstein structure in which the nonlinear part is represented by a memory-less tangent hyperbolic function (THF). Simulation results show that using the proposed PD along with the FXLMS algorithm increase the noise reduction performance significantly.

Wavelet network: online sequential extreme learning machine for nonlinear dynamic systems identification

A single hidden layer feedforward neural network (SLFN) with online sequential extreme learning machine (OSELM) algorithm has been introduced and applied in many regression problems successfully. However, using SLFN with OSELM as black-box for nonlinear system identification may lead to building models for the identified plant with inconsistency responses from control perspective. The reason can refer to the random initialization procedure of the SLFN hidden node parameters with OSELM algorithm. In this paper, a single hidden layer feedforward wavelet network (WN) is introduced with OSELM for nonlinear system identification aimed at getting better generalization performances by reducing the effect of a random initialization procedure.

Empirical modeling and control for spray drying of orange juice powder

Spray drying is a removal of moisture from liquid feed by breaking into droplets in a hot medium to convert into powder form. In order to ensure the product quality is at the desired specification, a good control system and good understanding on the dynamic behavior should be considered. The aims of this study are to develop empirical model of spray drying process and improve the process by implementation of PI controller. A nozzle atomizer spray dryer, Lab-Plant SD 05 Laboratory Scale Spray Dryer was used. The liquid feed was Sunquick Concentrated Orange Juice and DE 10–15 maltodextrin as the drying agent. The effects of inlet air temperature and maltodextrin concentration on final moisture content and outlet air temperature were investigated. From investigation, the effect of inlet air temperature on moisture content and outlet air temperature was greater than maltodextrin concentration. Thus, inlet air temperature was selected as manipulated variable. For modeling, the model obtained can be represented as first order process with time delay (FOPTD). In order to improve the process, the model obtained was used in simulation studies to determine the suitable tuning method by PI controller. The PI controllers were tuned by direct synthesis, min IAE method and Cohen-coon. From the observation, direct synthesis method is the most suitable tuning method for PI controller in spray drying process.

A review of SMD-PCB defects and detection algorithms

Detection and classification of defects on surface mount device printed circuit board (SMD-PCB) is an important requirement in electronic manufacturing process. This process which is primarily performed by automatic optical inspection (AOI) system ensures the functionality and quality of manufactured products. In this paper, the pattern recognition algorithms proposed in the literature for the inspection of defects using AOI are reviewed. The review focuses on segmentation algorithms, choice of features and feature extraction algorithms as well as the types of classifier and their relative classification performance. The review spans a 20 year period from 1990 to 2011. The results of the review suggest that solder joint defect is the type of defects mostly investigated and that the trend is moving towards combining the results of more than one classifier to improve classification accuracy and robustness.

Collision Prediction based Genetic Network Programming-Reinforcement Learning for Mobile Robot Navigation in Unknown Dynamic Environments

1921 Copyright c The Korean Institute of Electrical Engineers This is an Open-Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/ licenses/by-nc/3.0/) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. Collision Prediction based Genetic Network ProgrammingReinforcement Learning for Mobile Robot Navigation in Unknown Dynamic Environments

Performance comparison of THF-NLFXLMS and VFXLMS algorithms for Hammerstein NANC

Recently, THF-NLFXLMS algorithm was developed to compensate the nonlinearity encountered in nonlinear active noise control systems. Despite similar performance, this algorithm is more advantageous than the Nonlinear Filtered-X Least Mean Square (NLFXLMS) due to the use of tangential hyperbolic function (THF) instead of scaled error function (SEF) which allows the degree of nonlinearity to be modeled. In addition, the computational complexity is relatively small compared to other direct nonlinear adaptive algorithm like the Volterra filter. In this paper, the performance of THF-NLFXLMS algorithm for Hammerstein secondary path structure is quantified and compared with NLFXLMS and the Volterra Filtered-x Least Mean Squares (VFXLMS) algorithm of similar computational complexity. The results show that the THF-NLFXLMS algorithm has similar performance as NLFXLMS algorithm and outperforms 2nd order VFXLMS algorithm.

Performance comparison of LFXLMS, MOVFXLMS and THF-NLFXLMS algorithms for Hammerstein NANC

Filtered-X least mean square (FXLMS) algorithm is widely used in active noise control (ANC) systems when the secondary path is linear. However, the performance of FXLMS reduces when nonlinearity is present. Leaky FXLMS (LFXLMS) and minimum output variance FXLMS (MOVFXLMS) algorithms are effective in compensating the nonlinearity effects in nonlinear ANC (NANC). When using optimum leakage factors, these algorithms show close performance with benchmark nonlinear FXLMS (NLFXLMS) algorithm. In all three algorithms, the degree of nonlinearity must be known in advance and are usually assumed. In previous works, Tangential Hyperbolic Function NLFXLMS (THF-NLFXLMS) algorithm has been developed whereby the degree of nonlinearity is estimated using tangential hyperbolic function (THF). In this work, the performance of LFXLMS and MOVFXLMS based on optimum leakage factors calculated using the estimated degree of nonlinearity is compared with THF-NLFXLMS for Hammerstein structure. The results show that optimum MOVFXLMS performs similarly to optimum LFXLMS and THF-NLFXLMS.