Search

Our project mainly focuses work a lot being carried out in the field of distributed generation. For the electric grid, so Many distributed generation systems are being made. At this time when the non-renewable sources of energy such as oil, coal, etc are very fast disappearing, a study of distributed generation systems (DGs) and Establishing of such systems using renewable sources of energy becomes very important. we have seen several major problems will happen, when harmonic content of the current is flows into the grid, like this issue several issues are going to meet. When a Distribution Generation source is being linked to the grid. “The current being injected should have harmonic content conforming to standards such as IEEE 512-1992”[1].To limit the harmonic current injected into the grid. The hardware is tailor-made to be used for a solar PV panels based distributed generation system. This project deals mainly with Establishing the hardware for a grid connected inverter. This means that a proper plan of action should be present in the required system, special type of filters are made to abolish higher order harmonics, this is only the reason why filters would be less bulky and cheap. only the higher order harmonics are decreased by specially made filters, Lower order harmonics are attenuated by a technique called “an adaptive harmonic elimination technique (AHE). In our project we need to develop several blocks called the inverter hardware, design and implementation of closed loop control, transformer, the filters along with adaptive selective harmonic elimination scheme

In this study, a pavement surface texture measurement device is designed and fabricated based on active triangulation. The device acquires 3D surface texture data in a non-contact manner. The devic...

The goal of this paper is to compute the generating series of a closed-loop system when the plant is described in terms of a Chen-Fliess series and static output feedback is applied. The first step is to reconsider the so called Wiener-Fliess connection consisting of a Chen-Fliess series followed by a memoryless function. Of particular importance will be the contractive nature of this map, which is needed to show that the closed-loop system has a Chen-Fliess series representation. To explicitly compute the generating series, two Hopf algebras are needed, the existing output feedback Hopf algebra used to describe dynamic output feedback, and the Hopf algebra of the shuffle group. These two combinatorial structures are combined to compute what will be called the Wiener-Fliess feedback product. It will be shown that this product has a natural interpretation as a transformation group acting on the plant and preserves the relative degree of the plant.

At recent times, customer churn is an important activity in quickly developing industries such as telecom, banking, e-commerce, etc. Earlier studies revealed that the cost of getting a new customer is considerably higher than the cost of retaining the existing ones. Therefore, it becomes essential to predict the nature of customer churn for retaining the customers to a greater extent. The advent of deep learning (DL) models have begun to be employed for efficient CCP. This paper presents a new Adagrad Optimizer with Elephant Herding Optimization (EHO) based Hyper-parameter Tuned Bidirectional Long Short Term Memory (AG-EHO-BiLSTM) for CCP in Internet of Things (IoT) enabled Cloud Environment. The proposed AG-EHO-BiLSTM model initially acquires the customer data using its devices like smart phones, laptop, smart watch, etc. Next, the gathered data will be classified by the use of Bi-LSTM model, which determines the customers as churner or non-churner. The efficiency of the Bi-LSTM model can be increased through hyper parameter tuning techniques, namely Adagrad optimizer and EHO algorithm to optimally select the parameter values namely learning rate, number of hidden layer and epochs. The performance validation of the AG-EHO-BiLSTM model takes place on benchmark dataset and the simulation outcome reported the supremacy of the AG-EHO-BiLSTM model over the comparative methods.

The occurrence of motion blur along with defocus blur is a common phenomena in natural images. Usually, these blurs are spatially varying in nature for any general image and estimation of one type of blur is affected by presence of other. In this paper, we propose a novel method to estimate the concurrent defocus and motion blurs in a single image. Unlike the recent methods, which perform well only on simulated conditions or in presence of single type of blur, proposed method works well for real images as well as for compressed images. In this paper, we consider only commonly associated motion and defocus blurs for analysis. Decoupling of motion and defocus blur provides a fundamental tool that can be used for various analysis and applications.

Multi-sensor data fusion is advantageous while fusing data from heterogeneous range sensors, for scanning a scene containing both fine and coarse details. This paper presents a new multi-sensor range data fusion method with the aim to increase the descriptive contents of the entire generated surface model. First, a new training framework of the scanned range dataset to solve the relaxed Gaussian mixture model-based method by applying the convex relaxation technique is presented. The classification of the range data is based on a trained statistical model. In the data fusion experiments, a laser range sensor and Kinect (V1) are used. Based on the segmentation criterion, the range data fusion is performed by integration of the finer regions range data obtained from a laser range sensor with the coarser regions of the Kinect range data. The fused range information overcomes the weaknesses of the respective range sensors, i.e., the laser scanner is accurate but takes time while the Kinect is fast but not very accurate. The surface model of the fused range dataset generates a highly accurate, realistic surface model of the scene. The experimental results demonstrate robustness of the proposed approach.

One of the considerations, in the characterization of pavement surface, includes an accurate measurement of the pavement surface profile. With the advent of modern technologies, automated methods (such as laser scanning) are being preferred over manual methods (such as the sand patch test). The laser scanning measurement method is based on the principle of active triangulation, where a camera (laser detector), a laser emitter, and the illuminated laser spot form a triangle. However, due to obstruction from other nearby parts of the profile, this triangulation may not get completed for certain points of the surface profile, leading to occlusion of some data-points. In such a situation, generally, as per the current practice for pavement surfaces, interpolation is used to estimate the height of the occluded data points. In this study, an approach is suggested where two laser emitters are used (rather than one) to reduce this occlusion. Thus, even if the camera cannot see a point illuminated by one laser emitter, it may be able to capture the same point when it is illuminated by the other laser emitter. This minimizes (in some cases, altogether eliminates) occlusion. Experimental results are presented to support this approach.

Emerging Vehicle-to-Vehicle (V2V) technologies are expected to significantly contribute to the safety and growth of shared transportation provided challenges towards their deployment can be overcome. This paper focuses on one such challenge: characterizing the fraction of vehicles which have received a message, as a function of space and time, and operating under different traffic and communication conditions. V2V technologies bridge two infrastructures: communication and transportation. These infrastructures are interconnected and interdependent. To capture this inter-dependence, which may vary in time and space, we propose a new methodology for modeling information propagation between V2V-enabled vehicles. The model is based on a continuous-time Markov chain which is shown to converge, under appropriate conditions, to a set of clustered epidemiological differential equations. The fraction of vehicles which have received a message, as a function of space and time may be obtained as a solution of these differential equations, which can be solved efficiently, independently of the number of vehicles. Such characterizations can form the basis of assessing several attributes of V2V systems, some of which we demonstrate. The characterizations lend themselves to a variety of generalizations and capture various interdependencies between communication and mobility. As tests of the model we provide applications both in real-world settings using microscopic traffic traces and in postulated scenarios of outages and system perturbations: we find good model agreement with microscopic trajectory from two actual trajectory datasets, as well as a synthetic trajectory dataset generated from the origin/destination matrix.

Significant recent developments in 3D display technology have focused on techniques for converting 2D media into 3D. Depth map is an integral part of 2D-to-3D conversion. Combining multiple depth cues results in a more accurate depth map as it compensates for the errors caused by one depth cue as well as its absence by other depth cues. In this paper, we present a novel framework to generate a more accurate depth map for video using defocus and motion cues. The moving objects present in the scene are the source of errors in both defocus and motion-based depth map estimation. The proposed method rectifies these errors in the depth map by integrating defocus blur and motion cues. In addition, it also corrects the errors in other parts of depth map caused by inaccurate estimation of defocus blur and motion. Since the proposed integration approach relies on the characteristics of point spread functions of defocus and motion blur along with their relations to camera parameters, it is more accurate and reliable.

External border surveillance has become one of the most trending topics of research today. The problem of automating the detection in a big and Hercules terrains using WSN is one way to do. In first half of the paper, we have focussed on the chess queen-crossing strategy adopted by the intruder in order to cross the international border. In the other half, we have used milky-way deployment strategy of sensors to automate the process of detection to a particular zone and then forwarding that detected information to the nearest base station using homogeneous zone routing protocol. The relationship between the sensor detection and energy balancing is that in our work, the focus is on shifting the load of the task of the sensor node equally thereby reducing the burden of the nodes the energy spent in sensing, detecting and communication of the information regarding the movement of the direction of the intruder to a certain region and the base station deployed in that region is responsible for alerting the border action team to crack down on the unauthorized intruders especially during the night. To achieve this objective, we have classified the monitoring zone called border region into three zones and allocating three base stations for each zones. The simulations of work show that HZR protocol performs better in terms of network lifetime for this application.

The concept of relative degree plays an important role in nonlinear control theory. It provides, for example, a necessary and sufficient condition for the existence of a feedback linearizing control law for a single-input, single-output (SISO) input-affine nonlinear state space system. It also gives a sufficient condition under which a left inverse exists. In applications it is common for systems to be composed of smaller interconnected subsystems. It is known that various feedback structures preserve relative degree, but it is largely unknown how to compute the relative degree of interconnected systems using only their properties. So the goal of this paper is to determine the relative degree of two nonlinear control systems interconnected in a variety of different ways. A collection of illustrative examples is given.

Chalk and talk methods of teaching have been effective till the technological expansion has rooted up for student community. Finding some way to be distracted in a classroom, teaching and learning strategies needed an efficient approach for sharing/transfer of knowledge. An approach reported to be dynamic in design and supportive for 24/7 discussions with measures to monitor the progress periodically is proposed in this paper. DoDo takes an innovative concept of brute force method to prevent the missing of important points during a session. Exhaustion by mentioning every other standard in a concept of discussion enhances the level of learning. Accepting to the fact that not a 100 percent is prepared for sessions nor delivered in a classroom provides loopholes and justification for knowledge not to be shared. Assignments, tests and evaluations are supposed to identify the progress of students' skills rather rank them accordingly and categorize them. DoDo ensures the deliverance of concepts to every student irrespective of his/understanding capability, tests them with easier methodologies, provides a ground for practices and monitor them every now and then with a periodical report to explain the headway of indulgence.

There have been a number of techniques that have demonstrated the generation of multimedia data for one modality at a time using GANs, such as the ability to generate images, videos, and audio. However, so far, the task of multi-modal generation of data, specifically for audio and videos both, has not been sufficiently well-explored. Towards this, we propose a method that demonstrates that we are able to generate naturalistic samples of video and audio data by the joint correlated generation of audio and video modalities. The proposed method uses multiple discriminators to ensure that the audio, video, and the joint output are also indistinguishable from real-world samples. We present a dataset for this task and show that we are able to generate realistic samples. This method is validated using various standard metrics such as Inception Score, Frechet Inception Distance (FID) and through human evaluation., Comment: ICASSP 2021 (Accepted)

Compression of multimedia content is an important processing step and backbone of real life applications in terms of optimum resource utilization in transmission and storage. It is an established field of research with very little scope for further improvement in achieved compression through customary coding-based compression techniques. Consequently, non-customary compression methods have become an important area for future research. Based on the principle ‘Any information that can be restored can be compressed’, we propose a novel spectrum-based image compression technique to further reduce the data footprint with satisfactory quality metric for images. We first blur the image with a point spread function (PSF) determined using frequency content of the given image. Blurring increases the DC component in the image, which in turn gets further compressed compared with original image by DCT-based JPEG compression. To recover the image, we perform deconvolution using the known blur PSF. Results obtained show further improvement of $$20-30\%$$ in achieved compression with respect to original JPEG compressed image with satisfactory quality of recovered image.

A new learning methodology in terms of a discretization of a so-called Chen-Fliess series of a control affine nonlinear system was recently proposed, in part, for the purpose of systematically including system structure and expert knowledge into control strategies. The main objective of this paper is to appropriately embed this learning unit as a supporting predictive controller for power dynamical systems. In particular, an infinite bus system is used for the prototype design of a smart and active control policy to regulate voltage and frequency. It is demonstrated by simulation how a controller employing a Chen-Fliess learning unit can recover from a fault and address modeling mismatch.

At present times, Customer Churn Prediction (CCP) becomes a familiar research area and is mainly solved by the use of machine learning models. The advancements in Internet of Things (IoT) and cloud computing platform allows collecting data about the customers to carry out CCP. This paper introduces an optimal genetic algorithm (OGA) with support vector machine (SVM) model for CCP. Initially, the OGA is derived by the use of double chain-quantum-genetic-algorithm. Next, the derived OGA is applied for optimizing the parameters of SVM namely C and ϒ. The outcome of the OGA-SVM model is tested against a benchmark dataset from telecom industry. The experimental outcome indicated that the OGA-SVM shows excellent results by offering better sensitivity of 94.50, specificity of 66.06, accuracy of 90.27, F-Score of 94.30 and kappa value of 61.17 correspondingly.

Most of landmine detection robots proposed so far have been strongly restricted from locomotion inside the mine field because they cannot cross over the mine. So we have proposed a mine detection robot with hybrid locomotion, which can enter inside the minefield with low ground surface contact, which can cross over the mine instead of changing its path and scan landmines directly using Electro Magnetic Induction sensor. The hybrid locomotion proposed in the robot uses the frame walking technique and the conventional wheeled locomotion. The robot switches over the locomotion mechanism from wheeled to leg when mine is detected and vice versa with a lead screw mechanism. The leg locomotion is achieved by frame walking technique where the two frames translate with the help of lead screw mechanism. A purpose of adopting this combination is to evade anti-personnel landmines which are relatively smaller in comparison to their anti-tank landmine counterparts. The robot initially starts in wheeled mode and upon detection of metal, pulls in the frame walking algorithm. The robot also deploys an obstacle avoidance algorithm when working in wheel mode.

Information propagation in V2V-enabled transportation networks is highly influenced by both vehicle mobility and wireless communication. The mobility patterns and communication conditions are not only heterogeneous, but also vary both temporally and spatially. In particular, realistic traffic flow changes with time, exhibiting sharp time-triggered transitions, due to external factors such as traffic lights, unpredictable disruptions (e.g., accidents), and planned disruptions (e.g., road-block). More specifically, traffic signals cause traffic synchronization, due to vehicles stopping during the red phase, and starting almost simultaneously during the green phase, which fundamentally alters the dynamics of V2V message propagation in a complex manner. In this paper, we propose a mathematical framework, starting from a continuous-time Markov chain, that characterizes the fraction of vehicles that have received a message over time and space in an arbitrary road network even when the traffic flow exhibits sharp time-triggered transitions. Our framework can accommodate arbitrary traffic synchronization patterns corresponding for example to the presence of an arbitrary number of traffic signals. The stochastic model for V2V message flow converges to a set of differential equations as the number of vehicles increases. The analytical characterization lends itself to a fast computation regardless of the number of vehicles and traffic synchronization patterns, while vehicular network simulators can only realistically simulate small-scale transportation networks. We find that V2V simulations of a statistical model with traffic synchronization and simulation of communications applied on a synthetic traffic trace well match our model solution.

Measuring similarity between images is required in many multimedia applications such as retrieving perceptually similar images. The most widely used metric for measuring the distance between two signal vectors is mean square error. Even though MSE provides mathematical tractability, it is not suitable for detecting perceptual similarity between images. In this paper, we propose a new measure to calculate similarity between images called dictionary similarity measure. It determines the distance between two dictionaries learned with respective image features. Similarity measure using algorithm that we propose in this paper is calculated for various image datasets such as Vision Texture Database—MIT Media Lab, Corel-1000 image dataset, and Standard images (Mandrill, Boat, etc.). Further, we propose algorithms to retrieve perceptually similar images using maximum vote criterion and dictionary similarity measure. Similarity measure that we propose in the paper retrieves perceptually similar images in a computational efficient manner with nearly similar accuracy compared to some of the existing dictionary-based image retrieval techniques.

Although videos appear to be very high-dimensional in terms of duration × frame-rate × resolution, temporal smoothness constraints ensure that the intrinsic dimensionality for videos is much lower. In this paper, we use this idea for investigating Domain Adaptation (DA) in videos, an area that remains under-explored. An approach that has worked well for the image DA is based on the subspace modeling of the source and target domains, which works under the assumption that the two domains share a latent subspace where the domain shift can be reduced or eliminated. In this paper, first we extend three subspace based image DA techniques for human action recognition and then combine it with our proposed Eclectic Domain Mixing (EDM) approach to improve the effectiveness of the DA. Further, we use discrepancy measures such as Symmetrized KL Divergence and Target Density Around Source for empirical study of the proposed EDM approach. While, this work mainly focuses on Domain Adaptation in videos, for completeness of the study, we comprehensively evaluate our approach using both object and action datasets. In this paper, we have achieved consistent improvements over chosen baselines and obtained some state-of-the-art results for the datasets.

Watermarking is the process of embedding the particular information into the audio signal for managing the ownership copyrights through wireless network. During the watermarking process, the audio signal consumes high energy, conflicting problem of robustness and imperceptibility, signal to noise ratio, bit error rate and normalized correlation. To overcome these issues present in the audio watermarking process, novel wavelet decomposition and evolutionary algorithm is utilized. Initially the input information or message has been split into two and the spillted message is watermarked using the audio and the image in wireless network. The first half of the message is watermarked with the help of the image and the next half of the image is watermarked by audio. Initially the watermarked image is transferred into YIQ image, from the transferred image the scrambled image is generated with the help of the Hidden Markov tree counter let wavelet transform method for generating the watermarking image. Then the next part of information is watermarked by audio signal which is decomposed into various sub bands using the Multi-resolution complex dual tree wavelet method. From the decomposed audio signal, the message authentication code based watermarks have been embedded in the lower frequency coefficients. Then the embedded process is performed by using the Dead zone quantization process which is optimized with the help of the Fireflies algorithm for enhancing the quality of the watermarking process. Finally both watermarking process is embedded for improving the security to the information with efficient manner through wireless network. In addition the efficient watermarking process through wireless network reduces the various attacks while extracting the water marker. Thus the optimized wavelet and quantization process improves the image and audio watermarking through wireless network and efficiency of the proposed system is evaluated using the experimental results in terms of the Signal to noise ratio, normalized correlation and bit error rate.

This paper presents a comparative study of five different imaging techniques. These techniques play a very important role in detection of abnormalities in the masses by screening the part of our interest. Out of these, Tomosynthesis, which is a 3D imaging technique, allows detection of smaller tumors and smaller microcalcifications than the 2D images. Using this technique, screening can be done at a very small angle, which helps in acquiring the image of overlapping masses. Compared to all other techniques, Tomosynthesis requires less radiation doses.

Pavement surface characterisation usually consists of surface texture measurement and its consequent data analysis. Recently there has been significant development in automated and contactless surface characterisation techniques. This paper reviews some of these developments in connection with pavement surface characterisation.

The average life spans of light-emitting diodes (LEDs) used in organic light-emitting diode displays are negatively influenced by high current operation in short duty cycles. Since the time average brightness of a pixel LED is a function only of the time integral of the delivered current, it is desirable to achieve brightness by adopting pixel addressing schemes that employ long duty cycles. Conventional addressing methods which cannot operate more than one row of pixels at a time inherently have very short duty cycles. We solve the problem by proposing entire frame addressing, made possible by a new image factorization called monotonic-nonnegative matrix factorization (M-NNMF). Using M-NNMF, we develop a converging image series representation of an arbitrary nonnegative matrix (image). Each element of the series has unit rank, which allows it to drive an entire frame simultaneously. Each application of the M-NNMF algorithm produces a dominant unit rank component for the series, and a residue image of higher rank, the input to the next iteration. We drive the display with these unit rank components, called sub-frames, each for a duration proportional to its energy. The sum of sub-frames approximates the original image and provides the same visual effect, due to inherent perceptual integration. M-NNMF, more efficient and less time complex compared to existing NNMF algorithms, is our primary contribution in this paper and is likely to find applications in many other situations. We also obtain an even faster converging series with a randomized version of M-NNMF. The proposed approach of entire frame image display is demonstrated on a wide variety of images.

Objectives: The paper aims at reducing the power and energy deficit in various regions of India. Methods: A game theory based mathematical model performs the allocation called Shapley value, which does the fair allocation of power and energy in order to reduce the maximum deficit in a particular region. Findings: The simulation results confirm the distribution is fair enough in reducing the overall deficit percentage in all regions. Applications: The work can be integrated with the grid and further extensions to the next level are possible.

Active 3D vision is concerned with extracting 3D information from the visible surfaces of the scene. With the use of recent technological advances made in electronics, photonics and computer vision, it is now possible to develop a reliable, high-resolution and accurate 3D active range sensor. In this paper, first we present the development of Laser range sensor using only consumer-grade technologies to ensure affordable hardware cost and high angular resolution. In the designed Laser range sensor, a diode Laser projects the Laser line on the surface through a cylindrical lens while the camera captures the Laser line profile. Next, a new calibration technique of the range sensor is presented to acquire the range information based on relative pose between the Laser and the camera by taking single view of the scene. Experiments are carried out on real scene and efficient 3D reconstructed model of the objects has been obtained.

Text classification is an essential advance in characteristic dialect processing. It very well may be performed utilizing different classification algorithms. Hadoop Map Reduce is widely utilized in text classification to perform classification on colossal measure of text data. However, Map Reduce required a ton of time to perform the tasks thereby increasing latency and since the data is distributed over the cluster it builds time and thus reducing processing speed. Also, Hadoop utilizes long queue of code. Motivated by this, we propose a basic yet compelling machine learning method which uses Naive Bayes classifier for text data. In Machine Learning approach, the classifier is built automatically by learning the properties of categories from a set of predefined training data. Hence, it can process complex furthermore, multi assortmentinformation in dynamic situations. Here we propose a naive bayes classifier which scales directly with number of indicators and data points which can be used for both binary and multi-class classification problems. We implemented the presented schemes using Machine Learning tool. The experimental results demonstrate the performance improvement in the classification technique.

Defocus based Depth estimation has been widely applied for constructing 3D setup from 2D image(s), reconstructing 3D scenes and image refocusing. Using defocus enables us to infer depth information from a single image using visual clues which can be captured by a monocular camera. In this paper, we propose an application of Depth from Defocus to a novel, portable keyboard design. Our estimation technique is based on the concept that depth of the finger with respect to our camera and its defocus blur value is correlated, and a map can be obtained to detect the finger position accurately. We have utilised the near-focus region for our design, assuming that the closer an object is to our camera, more will be its defocus blur. The proposed keyboard can be integrated with smartphones, tablets and Personal Computers, and only requires printing on plain paper or projection on a flat surface. The detection approach involves tracking the finger’s position as the user types, measuring its defocus value when a key is pressed, and mapping the measured defocus together with a precalibrated relation between the defocus amount and the keyboard pattern. This is utilised to infer the finger’s depth, which, along with the azimuth position of the stroke, identifies the pressed key. Our minimalistic design only requires a monocular camera, and there is no need for any external hardware. This makes the proposed approach a cost-effective and feasible solution for a portable keyboard.

Artificial neural networks have traditionally been used to implement machine learning algorithms. There are, however, alternatives to these biologically inspired machine learning architectures that offer potentially lower complexity and stronger theoretical underpinnings. One such option in the context of control is based on using a generic input-output model known as a Chen-Fliess functional series. The main goal of the paper is to describe a specific architecture that can be used in the multivariable setting to combine both learning and model based control. It builds on recent work by the authors showing that a certain monoid structure underlies any recursive implementation of such a system. The method is demonstrated using a two-input, two-output Lotka-Volterra system.

Due to the need for higher quality depth data than possible with an individual range sensing approach nowadays, there has been a growing interest to develop an integrated depth sensing technique by fusion of different $3D$ acquisition approaches that are more precise than the individual devices. In this paper, a new unsupervised range data fusion method using distinct range sensors has been presented for the extraction of an accurate surface model. In the fusion method, the analysis of Kinect's depth data based on Haar wavelets is used to identify regions requiring finer scan by the Laser range sensor. The fused data illustrate the more accurate descriptive characteristic of the surface. The experimental results show a high quality reconstructed $3D$ model which validates the correctness of the real surfaces.

In this work, we proposed a novel depth adaptive tone mapping scheme for stereo HDR imaging and 3D display. We are interested in the case where different exposures are taken from different viewpoints. The scheme employed a new depth-adaptive cross-trilateral filter (DA-CTF) for recovering High Dynamic Range (HDR) images from multiple Low Dynamic Range (LDR) images captured at different exposure levels. Explicitly leveraging additional depth information in the tone mapping operation correctly identify global contrast change and detail visibility change by preserving the edges and reducing halo artifacts in the synthesized 3D views by depth-image-based rendering (DIBR) procedure. The experiments show that the proposed DA-CTF and DIBR scheme outperforms state-of-the-art operators in the enhanced depiction of tone mapped HDR stereo images on LDR displays.

Compression is an essential processing step for multimedia storage, distribution and transmission in terms of optimum resource utilization. Non-customary compression methods have become an important area for research due saturation of traditional coding based compression techniques. In this paper, we propose a novel spectrum based compression technique to further reduce the data footprint with satisfactory quality metric for the cases of video. Proposed method is based on principle that ‘Any information that can be restored, can be compressed’. We blur the individual frames with a given Point Spread Function (PSF) to increase the DC component and compress using standard MPEG coder. To recover the video, we perform frame by frame deconvolution using the known blur PSF. Results obtained show further improvement of 20 to 30% in achieved compression with respect to original MPEG compressed video with satisfactory quality of recovered input.

A hybrid control architecture for nonlinear dynamical systems is described which combines the advantages of model based control with those of real-time learning. The basic idea is to generate input-output data from an error system involving the plant and a proposed model. A discretized Chen-Fliess functional series is then identified from this data and used in conjunction with the model for predictive control. This method builds on the authors' previous work on model-free control of a single-input, single-output Lotka-Volterra system. The problem is revisited here, but now with the introduction of a model for the dynamics. The single-input, multiple-output version of the problem is also investigated as a way to enhance closed-loop performance.

Often high resolution (HR) RGB images generated by sparse sampling of the visible spectrum fail to produce differentiable modality for computer vision tasks, Hence computer vision tasks have to rely on gross structures in an image like corners, edges etc. instead of just the recorded reflectance of objects or materials at each pixel in a scene. In contrast to RGB, hyperspectral imaging allows pixels to record reflectance of the scene over multiple contiguous bands, which results in rich differentiable modalities. However, hyperspectral imaging, despite having a growing number of applications from agriculture, surveillance, mineralogy, food processing to eye care, is hitherto restricted to low spatial resolution imaging due to sensor hardware limitations. In this paper, we propose a hyperspectral super resolution technique to produce a high resolution (HR) hyperspectral image with a spectral support of 400–1020 nm from a low resolution (LR) hyperspectral image of the same spectral support and a high resolution multispectral (RGB) image with reduced spectral support of 400–700 nm. In the first step, we generate a HR hyperspectral prior by estimating HR hyperspectral band images in the spectral support of 400–700 nm by detail transfer and alternating iterative minimization. In the next step, we use the generated HR prior to further estimate the HR hyperspectral images for 710–1020 nm bands by learning a non-negative dictionary of reflectance spectra signatures of all the materials present in the scene from the LR hyperspectral image with spectral support in 400–1020 nm. With the estimated HR hyperspectral prior and learned dictionary, we predict the non-negative sparse codes for HR hyperspectral band images in the band of 710–1020 nm.

The work gives a novel evolutionary technique for the re-configuration test IEEE systems. The framework applied for optimization is Symbiotic Organism Search Algorithm (SOSA). The aim is to find optimal reconfiguration and to improve the active loss in the consumer side. This approach is examined on 16 and 33 IEEE test circuits. The results shows a valid improvement of active losses. The time required for execution is less when compared to other approaches.

A formal power series over a set of noncommuting indeterminants using iterated integrals as the coefficients is called a Chen series, named after the mathematician K.-T. Chen. The first goal of this paper is to give a brief overview of Chen series and their algebraic structures as a kind of reference point. The second goal is to describe its discrete-time analogue in detail and then apply the concept in two problems, the time discretization problem for nonlinear control systems and the machine learning problem for dynamical systems.

Content-based video retrieval and video synopsis are generally considered as two different areas. In this paper, we present an efficient approach for video retrieval based on the perceptual synopsis database of the videos. Video synopsis encapsulates an overview of a shot in a single frame. This is the first time video synopsis is used for video indexing providing the user an intuitive link for accessing actions in the video. We propose an enhanced synopsis called meta synopsis for the video database index, which will contain all essential information for retrieval. Various information such as background of a scene, motion trajectory of the foreground objects, color, texture, and mutual information in the synopsis database will empower us to retrieve relevant video content from huge video databases. Experiments were conducted on the OVP, BBC Motion Gallery, TRECVID data set, and other videos. Instead of using key frames as the query frames, the method accepts any arbitrary query frames. The experimental results illustrate that our proposed method can accurately identify a pertinent video from huge video databases.

Studies on the propagation of malware in mobile networks have revealed that the spread of malware can be highly inhomogeneous. Platform diversity, contact list utilization by the malware, clustering in the network structure, etc. can also lead to differing spreading rates. In this paper, a general formal framework is proposed for leveraging such heterogeneity to derive optimal patching policies that attain the minimum aggregate cost due to the spread of malware and the surcharge of patching. Using Pontryagin's Maximum Principle for a stratified epidemic model, it is analytically proven that in the mean-field deterministic regime, optimal patch disseminations are simple single-threshold policies. Through numerical simulations, the behavior of optimal patching policies is investigated in sample topologies and their advantages are demonstrated., Comment: 18 pages, 6 figures

Rescue missions are very critical in saving the lives of many people. The people who involve themselves in the same risk their lives in order to save the lives of the victims. In such cases, tracking a person in case of dense areas can be of huge difficulty as GPS signal will not be in a position to penetrate through the buildings. They have the chance of getting reflected back, thereby leading to erroneous prediction of the location of the person. It would be handy to develop a wearable that can help in tracking the person. This can be of great use during rescue operations in forests. JARVIS is the device that we have developed, which is a personalized monitoring system to track the person in case if the place has no access to GPS using an Inertial Measurement Unit (IMU). The IMU sensor has accelerometer, gyrometer and magnetometer which can be used to estimate the route. Pedestrian Dead Reckoning algorithm and Zero Velocity Update algorithm have been used in order to find the position of a person relative to the initial position that can be fixed. We estimate the position of the person with an accuracy of 50 cm. Moreover, it becomes essential to predict the state of a person who is involved in rescue missions as the life of the person is at huge risk. Hence, the activity of a person is being recognized using Machine learning in real-time and is sent to a server. In addition to the above mentioned features, we use pulse sensor and temperature sensor in order to track the health condition of a person so that we get information about the condition of the person so that first aid can be given in case of emergency. This can be used in various rescue missions such as fire rescue operations and military operations where the person who is involved in rescue mission needs to be monitored so that effective first aid can be given to him promptly in case of unexpected body state of the person. Thus, JARVIS can be an effective device in such situations.

High-bandwidth Video Bus is a video interface protocol defined for high bandwidth, low latency, and uncompressed digital video transmission in avionics systems. This protocol is used for interfacing aircraft system(s) into the cockpit display. Current day technology realizes the systems or Line Replaceable Units (LRU) using Field-Programmable Gate Array (FPGA) based Intellectual Property (IP) Cores for communication between two systems. However, these communication protocols are realized using FPGA and as part of development, the Peripheral Component Interconnect (PCI) express based FPGA embedded modules are used for realization. These embedded modules will be interfaced to the host computers to transmit the data as similar to the communication on aircraft for real LRU. Hence host computers need to be integrated with required GUI application which shall communicate with PCI express bus connecting the FPGA based embedded module. This paper describes the development of the GUI application program for PCI express data simulator using device drivers generated by JUNGO (WinDriver) tool for memory read/write transaction(s). The paper also presents the integration, testing and data transfer for discrete and continuous data over PCI express from the host machine to the embedded FPGA-based module.

The problem of domain adaptation has been extensively studied for object classification task. However, this problem has not been as well studied for recognizing actions. While, object recognition is well understood, the diverse variety of videos in action recognition make the task of addressing domain shift to be more challenging. We address this problem by proposing a new novel adaptation technique that we term as unsupervised deep action domain adaptation (U-DADA). The main concept that we propose is that of explicitly modeling density based adaptation and using them while adapting domains for recognizing actions. We show that these techniques work well both for domain adaptation through adversarial learning to obtain invariant features or explicitly reducing the domain shift between distributions. The method is shown to work well using existing benchmark datasets such as UCF50, UCF101, HMDB51 and Olympic Sports. As a pioneering effort in the area of deep action adaptation, we are presenting several benchmark results and techniques that could serve as baselines to guide future research in this area.

Representation learning is the central role for all natural language processing task. Bag of Words method lacks in semantics and word order, hence word embedding model Word2Vec is used to capture the word semantics. But for morphological rich language, the vector representation would be noisy due to polysemy. To address these problems, Bag of Concepts is introduced to capture association between the words in the documents and forms concept cluster. Sometimes, Bag of Concepts representations may ignore the syntax for large amount of data. Hence a novel ensemble representation learning method for document classification is proposed by combining the word2vec and Bag of Concepts model to tackle the above mentioned problems. Extensive results on the Reuter datasets, show that the proposed model for document classification outperforms the baseline model in terms of F1 score.

This paper presents an algorithm for noise removal from digital image, based on stochastic noise pattern. We propose spatially localized stochastic resonance for image denoising implemented using Suprathreshold Stochastic Resonance (SSR) and Dynamic Stochastic Resonance (DSR). The approach has been tested on 300 images. For SSR implementation, input noisy image is subjected to independent additive white Gaussian noise of different standard deviation, the output image corresponding to individual noise standard deviation, summed and averaged, to get the denoised image. For DSR implementation, singular values are used for noise suppression in digital images. Key issue of our work is the reconstruction of the input noisy image by stochastic noise pattern using SSR and DSR that reflect better features of the image. The results of these are quantified appropriately through PSNR (peak signal-to-noise ratio) and visualization of an output image. The localized implementation was tested on a dataset of 300 images and when compared with the conventional techniques, our approach is found to give marginally better noise reduction in most of the cases. The average PSNR improvement is considerably improved.

Motion planning of rovers in rough terrains involves two parts of finding a safe path from an initial point to a goal point and also satisfying the path constraints (velocity, wheel torques, etc.) of the rover for traversing the path. In this paper, we propose a new motion planning algorithm on rough terrain for a 6 wheel rover with 10 DOF (degrees of freedom), by introducing a gradient function in the conventional potential field method. The new potential field function proposed consists of an attractive force, repulsive force, tangential force and a gradient force. The gradient force is a function of the roll, pitch and yaw angles of the rover at a particular location on the terrain. The roll, pitch and yaw angles are derived from the kinematic model of the rover. This additional force component ensures that the rover does not go over very high gradients and results in a safe path. Weights are assigned to the various components of the potential field function and the weights are optimized using genetic algorithms to get an optimal path that satisfies the path constraints via a cost function. The kinematic model of the rover is also derived that gives the wheel velocity ratio as it traverses different gradients. Quasi static force analysis ensures stability of the rover and prevents wheel slip. In order to compare different paths, four different objective functions are evaluated each considering energy, wheel slip, traction and length of the path. A comparison is also made between the conventional 2D potential field method and the newly proposed 3D potential field method. Simulation and experimental results show the usefulness of the new method for generating paths in rough terrains. Proposed a new potential field method for rough terrain path planning for a rover.A gradient function is introduced in the conventional potential field method.The gradient function depends on the roll, pitch and yaw angles of the rover.Weights of potential field function are optimized by using GA.Results prove that the new method is superior to conventional potential field method.