Showing papers by "Marizan Mubin published in 2017"

Improving EEG signal peak detection using feature weight learning of a neural network with random weights for eye event-related applications

Abstract: The optimization of peak detection algorithms for electroencephalogram (EEG) signal analysis is an ongoing project; previously existing algorithms have been used with different models to detect EEG peaks in various applications. However, none of the existing techniques perform adequately in eye event-related applications. Therefore, we aimed to develop a general procedure for eye event-related applications based on feature weight learning (FWL), through the use of a neural network with random weights (NNRW) as the classifier. The FWL is performed using a particle swarm optimization algorithm, applied to the well-studied Dumpala, Acir, Liu and Dingle peak detection models, where the associated features are considered as inputs to the NNRW with and without FWL. The combination of all the associated features from the four models is also considered, as a comprehensive model for validation purposes. Real EEG data recorded from two channels of 20 healthy volunteers were used to perform the model simulations. The data set consisted of 40 peaks arising in the frontal eye field in association with a change of horizontal eye gaze direction. It was found that the NNRW in conjunction with FWL has better performance than NNRW alone for all four peak detection models, of which the Dingle model gave the highest performance, with 74% accuracy.

Attitude control of drones by back-stepping controller

Abstract: The drones are used for many purposes such as structure inspections, pesticide sprayings, photographys, load carriages and investigations in the world. The drones have a week point that it is difficult to fly them in the strong wind. However, it is an important task to fly in the strong wind for the disaster relieves. In this paper, we show the modeling and the attitude control system design of a drone by using a back-stepping controller, and good performances are shown in simulations.

Adaptive switching gravitational search algorithm: an attempt to improve diversity of gravitational search algorithm through its iteration strategy

Abstract: An adaptive gravitational search algorithm (GSA) that switches between synchronous and asynchronous update is presented in this work. The proposed adaptive switching synchronous–asynchronous GSA (ASw-GSA) improves GSA through manipulation of its iteration strategy. The iteration strategy is switched from synchronous to asynchronous update and vice versa. The switching is conducted so that the population is adaptively switched between convergence and divergence. Synchronous update allows convergence, while switching to asynchronous update causes disruption to the population’s convergence. The ASw-GSA agents switch their iteration strategy when the best found solution is not improved after a period of time. The period is based on a switching threshold. The threshold determines how soon is the switching, and also the frequency of switching in ASw-GSA. ASw-GSA has been comprehensively evaluated based on CEC2014’s benchmark functions. The effect of the switching threshold has been studied and it is found that, in comparison with multiple and early switches, one-time switching towards the end of the search is better and substantially enhances the performance of ASw-GSA. The proposed ASw-GSA is also compared to original GSA, particle swarm optimization (PSO), genetic algorithm (GA), bat-inspired algorithm (BA) and grey wolf optimizer (GWO). The statistical analysis results show that ASw-GSA performs significantly better than GA and BA and as well as PSO, the original GSA and GWO.12

Transitional particle swarm optimization

Abstract: A new variation of particle swarm optimization (PSO) termed as transitional PSO (T-PSO) is proposed here. T-PSO attempts to improve PSO via its iteration strategy. Traditionally, PSO adopts either the synchronous or the asynchronous iteration strategy. Both of these iteration strategies have their own strengths and weaknesses. The synchronous strategy has reputation of better exploitation while asynchronous strategy is stronger in exploration. The particles of T-PSO start with asynchronous update to encourage more exploration at the start of the search. If no better solution is found for a number of iteration, the iteration strategy is changed to synchronous update to allow fine tuning by the particles. The results show that T-PSO is ranked better than the traditional PSOs.