Entropies for detection of epilepsy in EEG

The electroencephalogram (EEG) is a representative signal containing information about the condition of the brain. The shape of the wave may contain useful information about the state of the brain. However, the human observer cannot directly monitor these subtle details. Besides, since bio-signals are highly subjective, the symptoms may appear at random in the time scale. Therefore, the EEG signal parameters, extracted and analyzed using computers, are highly useful in diagnostics. The aim of this work is to compare the different entropy estimators when applied to EEG data from normal and epileptic subjects. The results obtained indicate that entropy estimators can distinguish normal and epileptic EEG data with more than 95% confidence (using t-test). The classification ability of the entropy measures is tested using ANFIS classifier. The results are promising and a classification accuracy of about 90% is achieved.     

Non-linear analysis of EEG signals at various sleep stages

Application of non-linear dynamics methods to the physiological sciences demonstrated that non-linear models are useful for understanding complex physiological phenomena such as abrupt transitions and chaotic behavior. Sleep stages and sustained fluctuations of autonomic functions such as temperature, blood pressure, electroencephalogram (EEG), etc., can be described as a chaotic process. The EEG signals are highly subjective and the information about the various states may appear at random in the time scale. Therefore, EEG signal parameters, extracted and analyzed using computers, are highly useful in diagnostics. The sleep data analysis is carried out using non-linear parameters: correlation dimension, fractal dimension, largest Lyapunov entropy, approximate entropy, Hurst exponent, phase space plot and recurrence plots. These non-linear parameters quantify the cortical function at different sleep stages and the results are tabulated.     

Multiperson interaction recognition in images: A body keypoint based feature image analysis

Most interaction recognition approaches have been limited to single‐person action classification in videos. However, for still images where motion information is not available, the task becomes more complex. Aiming to this point, we propose an approach for multiperson human interaction recognition in images with keypoint‐based feature image analysis. Proposed method is a three‐stage framework. In the first stage, we propose feature‐based neural network (FCNN) for action recognition trained with feature images. Feature images are body features, that is, effective distances between a set of body part pairs and angular relation between body part triplets, rearranged in 2D gray‐scale image to learn effective representation of complex actions. In the later stage, we propose a voting‐based method for direction encoding to anticipate probable motion in steady images. Finally, our multiperson interaction recognition algorithm identifies which human pairs are interacting with each other using an interaction parameter. We evaluate our approach on two real‐world data sets, that is, UT‐interaction and SBU kinect interaction. The empirical experiments show that results are better than the state‐of‐the‐art methods with recognition accuracy of 95.83% on UT‐I set 1, 92.5% on UT‐I set 2, and 94.28% on SBU clean data set.     

Promoting Software Reuse Using Self Organizing Maps

Reusability of software, regardless of its utilizing technique, is widely believed to be a promising means for improving software productivity and reliability. However it is not practiced adequately due to the lack of techniques that facilitate the locating of reusable components that are functionally close. In this paper we apply Kohonen's Self-Organizing Maps to develop an approach for promoting Software Reuse. We look at the details of how Self-Organization can arrange and regularize data from the original pattern space into a topology preserving map. We describe a practical implementation of the SOM methodology for Software Reuse using a database of UNIX commands. And finally we briefly present our proposed Software Reuse Methodology.     

Direct Numerical Simulation of pulsed jets-in-crossflow

Direct Numerical Simulations of a jet with a passive scalar injected vertically into a crossflow (velocity ratio = 6) is performed at a jet Reynolds number of 5000. The role of sinusoidal forcing of the jet on the dynamics of the flow structures, and on the jet trajectory and jet spreading is examined. Sinusoidal excitations selected are at non-dimensional frequencies of 0.2, 0.4, and 0.6. For the unforced jet, shear-layer vortices on the leading edge of the jet have a preferred mode frequency of around 0.35. With forcing, the dominant frequency in the near field of the jet is the forcing frequency, but further downstream, vortex interactions/mergings lead to the growth of the subharmonic modes. For a forcing frequency of 0.2, the jet bifurcates in the vertical plane; at a forcing frequency of 0.4, the jet trifurcates into three jet-streams in the vertical plane; and, at a forcing frequency of 0.6, the jet bifurcates in the horizontal plane. The largest vertical penetration is at a forcing frequency of 0.4, while the largest horizontal spreading occurs at a non-dimensional forcing frequency of 0.6. Wake vortices, with a U-loop structure, are seen for all cases except at a forcing frequency of 0.6, where they are completely suppressed. The U-loop structure is asymmetric for the unforced and 0.2 forcing frequency case, and is consistent with the earlier experimental observations for unpulsed jets-in-crossflow. Through particle visualizations, a mechanism for the development of the wake vortices is presented in the paper. Mean statistics on isosurface contours are also presented, and asymmetry in the mean counter-rotating kidney pair vortex structures are also seen for the unforced and 0.2 forcing cases. The results of this study indicate the potential of using external forcing as a potential control strategy for controlling the jet penetration and mixing with the crossflow in either the vertical plane or the horizontal plane.     

A Case‐Based Reasoning system for complex medical diagnosis

A Case‐Based Reasoning (CBR) system for medical diagnosis mimics the way doctors make a diagnosis. Given a new case, its accuracy in practice depends on successful retrieval of similar cases. CBR systems have had some success in dealing with simple diseases because of the robustness of their case base. However, their diagnostic accuracy suffers when dealing with complex diseases particularly those that involve multiple domains in medicine. An example of such a condition is Premenstrual syndrome (PMS) as it falls under both gynaecology and psychiatry. To address this issue, the paper proposes a CBR‐based expert system that uses the K‐nearest neighbour (KNN) algorithm to search k similar cases based on the Euclidean distance measure. The novelty of the system is in the design of a flexible auto‐set tolerance (T), which serves as a threshold to extract cases for which similarities are greater than the assigned value of T. A prototype software tool with a menu‐driven Graphical User Interface (GUI) has been developed for case input, analysis of results, and case adaptation within the system. Finally, the performance of the tool has been checked on a set of real‐world PMS cases.     

A Series-elastic Robot for Back-pain Rehabilitation

This paper addresses the robot-assisted rehabilitation of back pain, an epidemic health problem affecting a large portion of the population. The design is composed of two springs in series connected to an end-effector via a pair of antagonistic cables. The spring and cable arrangement forms an elastic coupling from the actuator to the output shaft. An input-output torque model of the series-elastic mechanism is established and studied numerically. The study also illustrates the variation of the mechanism’s effective stiffness by changing the springs’ position. In addition, we built a prototype of the robotic mechanism and design experiments with a robotic manipulator to experimentally investigate its dynamic characteristics. The experimental results confirm the predicted elasticity between the input motion and the output torque at the end-effector. We also observe an agreement between the data generated by the torque model and data collected from the experiments. An experiment with a full-scale robot and a human subject is carried out to investigate the human-robot interaction and the mechanism behavior.

     

Limits of compensation in a failed antenna array

In large antenna arrays, the possibility of occurrence of faults in some of the radiating elements cannot be precluded at all times. In such situations, the radiation pattern of the array gets distorted, mostly with an increase in sidelobe level and decrease in gain. Although it is not possible to restore the pattern fully by rearranging the excitations of the functioning elements, compensation methods have been reported in the literature for restoring one performance parameter of the array and making a trade‐off on some other parameter. In this article, we have made a study on the tolerance level of this compensation process. One part of the study deals with the thinning in the failed array, that is, to find a limit on the minimum number of functioning elements of the array that can restore the digital beamforming of the failed array. The second part of study deals with finding the maximum number of element failures that can be compensated. The study was carried out by optimizing the amplitude excitations of the failed array. Instead of classical optimization techniques, particle swarm optimization was used for the compensation process. © 2014 Wiley Periodicals, Inc. Int J RF and Microwave CAE 24:635–645, 2014.