7940 research outputs found
Sort by
Escalation of Modified Condition/Decision Coverage for Object-Oriented Programs
No full textSoftware testing is acknowledged as quite possibly the best procedure to guarantee the quality of software. Particularly, white-box testing found to be a successful and effective technique to generate test cases. In the past few years, several automatic test generation techniques were proposed, among which ‘‘Symbolic Execution” (SE), ‘‘Concolic Execution” (CE) and ‘‘Dynamic Symbolic Execution” (DSE) are the state-of-art techniques. Usually, these techniques traverse different possible program execution paths based on a given coverage criterion. During this process they collect the generated symbolic constraints at conditional statements and explore the possible execution paths. The constraint solver produces an input subspace that is used to generate new test cases for finding a path on the execution tree. Some of the existing test generation tools are CUTE, CATG, KLEE, Jalangi etc. However, literature says that these tools as an individual are not competent enough for generating effective test data that could lead to increase ‘‘Modified Condition/Decision Coverage” (MC/DC) score. This is because of certain limitations associated with the testing tools. So, these tools are plugged with some additional approaches and techniques to improve the code coverage like MC/DC. Different techniques have been reported in this area to improve MC/DC score. Some of the existing work used source code transformation techniques, program repair techniques, code optimization techniques and test data integration based on hybrid techniques for achieving high MC/DC score. But, looking at the present day results, it is expected that still it needs more attention to improve MC/DC score. So, to visualize and extend the current trend of research, in our work jCUTE is used to generate test data. The primary objective of our work is to improve the test data effectiveness, which in turn helps increase MC/DC score. These effective test data are generated through different approaches that are proposed in the contributing chapters. First contribution in this thesis is to propose a hybrid mechanism to combine concolic testing and search based testing together to produce more number of effective test cases that can increase MC/DC score. Here, jCUTE is used for concolic testing and Evosuite is used for search based testing. Both tools produce test cases and then we combine these test suites to make a single test suite. This final test suite is minimized to remove all duplicate and ineffective test cases. The experimental results show that this approach achieves high MC/DC than the traditional individual techniques. Then another integrated black box approach is proposed to improve MC/DC score which shows better code coverage. The generated test cases using concolic testing are not sufficient alone to meet the required coverage level for MC/DC. Thus, our second approach combines boundary value analysis technique with concolic test generation process to populate additional test cases. This add-on technique generates more number of effective test cases by populating the original set of test cases. Now, the updated test suite is minimized so as to reduce the testing cost. This technique is compared with some of the existing techniques and found to be much more effective than the existing works. Next, a technique is proposed to prioritize the generated test cases using two different parameters such as ‘‘contribution index” (CI) and ‘‘fault exposing potential” (FEP). The MC/DC coverage is measured using the proposed ‘‘Java MC/DC Analyzer” (JMA) tool. Similarly, FEP is also auto calculated using spectral analysis of predicates present in the input program. The effectiveness of the prioritized test cases is validated using a metric called‘‘Prioritization Index” (PI), which is calculated by adding CI and FEP together. PI helps rank the test cases for prioritization. The inference from this work indicates that the code coverage should be used only as a factor for test case prioritization. Therefore, the researchers should focus on multiple parameters or factors. This research work proposes different approaches for improving test case generation process and MC/DC score. These approaches help automate MC/DC computation process. These approaches are also validated for handling small to large benchmark Java programs under different scenarios and achieve significant improvement over the traditional techniques. The final contributing work in this thesis is to propose a source code transformation technique to generate effective test cases to enhance MC/DC. This source code transformation technique is proposed using concolic testing that helps achieve high MC/DC. This code transformation technique inserts code blocks into the program without violating the program’s outcome. This helps create several new branches in the test program. This is also called program restructuring. It may be noted that this process does not modify the original code, but inserts additional code blocks in the program. These new branches help explore more paths while executing the program. This increases MC/DC score. The proposed code transformation technique is named ‘‘Pattern Based code transformation”(PBCT). PBCT uses Boolean expression (predicate) for evaluation to traverse all the possible paths during program execution. The specific boolean patterns are used to help generate effective test cases to increase MC/DC score. For this process, an in-house developed tool is used. This approach is experimented with several benchmark Java programs taken from different data repositories. These input programs are executed having minimum 4 up to 2866 number of conditions to generate effective test cases. The effectiveness of our technique is compared with the existing literature. This comparison shows that our technique achieves considerably better coverage than the existing technique
Rainfall Characteristics Over The Indian Subcontinent And The Influence Of The Madden Julian Oscillation
No full textUnderstanding the influence of the Madden Julian Oscillation (MJO) on various rainfall characteristics is intriguing and important over the Indian subcontinent. Previous studies have documented the various aspects of rainfall, but a comprehensive understanding of diurnal cycle of rainfall (DCR) and extreme rainfall events (ERE) during MJO phases is still warranted across the seasons. The rainfall occurs in conjunction with the primary climatic modes (e.g. ENSO, MJO, and IOD) might invigorate or enervate the rainfall, which has many societal repercussions. Therefore, an in-depth understanding of the influence of primary climate modes on rainfall is essential to mitigate the vast impacts. With this rationale, the dissertation aims to quantify the scale interaction of the MJO during the diurnal and ERE over the Indian subcontinent. The Tropical Rainfall Measuring Mission precipitation data were used to manifest DCR and ERE. The European Centre for Medium-Range Weather Forecasts and National Centers for Environmental Prediction and the National Center for Atmospheric Research reanalysis data sets were used to investigate large-scale circulations associated with the rainfall. In order to delineate the MJO phases, the Real-time Multivariate MJO index was employed. The phases of the MJO are classified into active and suppressed, depends on the location of active convective centers over the oceanic regions of Indo-Pacific. The rainfall frequency percentage and the rainfall contribution index were used to understand the diurnal cycle of rainfall. The peak over the threshold, changes in probability, logistic regression models, standardized precipitation index and composite analysis were used to understand the statistical and atmospheric circulation during extreme rainfall events.The perusal of results reveals that the positive rainfall anomalies and enhanced rainfall peak during the active phases of the MJO. However, negative rainfall anomalies and declined rainfall peaks were evident during the MJO suppressed phases. The pattern is consistent across the seasons and prominently evident over the Bay of Bengal, equatorial Indian Ocean and east coastal regions of India. The diurnal cycle of rainfall shows the mid-late afternoon peak over the land regions and the late evening to early morning peak over the oceanic regions. In addition, during the Indian Summer Monsoon (ISM), the nocturnal peak was evident over the Himalayan mountains and Western Ghats regions. Similarly, the secondary rainfall peak during the daytime was evident over the central Bay of Bengal regions. Importantly, this study identified the presence of seashore and offshore regimes during the DCR over the Arabian Sea and Bay of Bengal coastal regions Similarly, the analysis of sub-daily rainfall extremes reveals that ERE intensity was higher than the daily ERE during pre-monsoon and ISM. The frequency percentage index of hourly rainfall events during different octets shows the mid to late afternoon (MLA) and late evening-early morning (LE-EM) peaks during ISM. This signifies that the sub-daily extremes were phase-locked with the diurnal cycle, notably during the ISM. The probability of detection analysis affirms the above results. Such as the daytime extremes were prominent over the land, and nocturnal extremes were occurring over coastal and oceanic regions. The investigation of synoptic circulations shows that the nocturnal extremes are stronger over the CIR and the WCWG regions, whereas the daytime extremes are stronger over the SIP. Indeed, the synoptic circulations and mechanisms, the nocturnal extremes are governed by air pump of the second kind, and daytime follows the inertial oscillation theory during the ISM. Akin to this, the analysis of ERE using the peak-over threshold method reveals the highest intensity of events prevails during the ISM over the Western Ghats, Northeast India and monsoon core regions. During the post-monsoon seasons in the eastern coastal regions of India, the intensity of rainfall events was higher. The influence of MJO was invariably evident in the distribution of extreme rainfall events. For instance, the frequency and rainfall contribution of the ERE were found higher during the active phases of the MJO. The south of 20o N was identified as rainfall hotspot regions, where active MJO influences were predominant. The logistic regression and cumulative probability changes affirm the above results. This study also aims to understand the precursors, mechanisms and synoptic-scale circulations manifested in ERE. The Indian subcontinent is divided into three homogenous regions such as Southern Indian Peninsula (SIP), Western Coast and the Western Ghats regions (WCWG) and the central Indian regions (CIR) for further analysis. The standardized precipitation index (SPI) was computed using the gamma distribution function to separately identify the top hundred SPI events over each region. The analysis reveals that the presence of meridional dipole plays a vital role in occurrences of ERE in the ISM over the SIP and WCWG regions. The anomalous high over the head Bay of Bengal and anomalous low over the south of 14o N ensure a conducive environment for the ERE. Notably, ERE were prominent over the break–active transition phases. Over the CIR, the intricate interaction between the oscillating monsoon trough, westward-moving monsoon depressions in the presence of secondary cyclonic vortices causes widespread extremes. During the post-monsoon seasons, the formation of tropical low-pressure systems over the North Indian Ocean (NIO) plays a vital role in the occurrences of ERE across the regions. The formation of mesoscale convective systems over zones of the dry-moist environment leads to rainfall extremes in the pre-monsoon season. In addition, the southward intriguing of subtropical westerly jet and northward propagating pre-monsoon organized convection from Indian Ocean regions causes the ERE. The large-scale circulations during the active and suppressed phases indicate that the contextual synoptic background provided by the MJO might facilitate or impede the occurrence of rainfall extremes over the homogenous regions during ISM. The passage of active MJO plays a prominent role in the genesis and intensification of the tropical low-pressure systems over the NIO. Overall, the present study enhances the existing knowledge on diurnal and extreme rainfall characteristics and their association with the MJO over the Indian subcontinent. In addition, in the current warming scenario, the evidence suggests that the frequency of the MJO will escalate; therefore, this study has practical implications in the seasonal to sub-seasonal prediction, flood forecasting, water resources management and mitigation and to validate advanced high-resolution ocean-atmospheric coupled models
Design and Development of Hybrid FSO/RF Communication System with Auto-Tracking Mechanism
No full textOptical wireless communication (OWC) has an enormous potential to support massive data transmission requirements. In an outdoor environment, OWC with long-range communication is referred to as free space optics (FSO) in the literature. FSO is one of the technologies which supports high bandwidth, unlicensed spectrum, high security, immune to interference and ease of installation. These attractive features of FSO provide a viable solution to the last mile problem in broadband wireless transmission. FSO communication remains sensitive to environmental conditions, scintillation and pointing errors despite many advantages. The design of FSO communication must consider all the above channel impairments. This research work focuses on designing reliable and available FSO communication under different channel conditions. The design of a hybrid FSO/RF communication with an auto-tracking system defines the scope of this work. Approaches adopted for the design of a hybrid FSO/RF system are enumerated below: • Initially, the performance of FSO communication in various atmospheric conditions and pointing losses are measured. In this regard, different weather conditions (i.e., rain, fog, and snow), atmospheric turbulence, and geometrical losses are taken into account to evaluate the FSO system performance. Also, an experimental testbed for FSO communication is designed and implemented with an indoor atmospheric chamber to replicate the weather conditions. An image under different foggy and atmospheric turbulence conditions is transmitted to evaluate the performance. • In FSO communication, fog, atmospheric turbulence, and pointing errors are the major bottlenecks that degrade the performance significantly. On the other hand, RF communication is less prone to the above issues. A hybrid FSO/RF is designed and implemented with an auto-tracking system. An efficient machine learning (ML) aided switching mechanism is proposed for selecting the appropriate communication link based on the current weather conditions. The ML model is trained with different weather conditions to estimate the link margin (LM) of FSO communication. A switching decision is taken based on the LM estimation. • Even under clear weather conditions, the performance of FSO communication is primarily dependent on atmospheric fading and a strict line-of-sight (LoS) condition,i.e., pointing errors. Therefore, a statistical model for intensity fluctuation due to the atmospheric turbulence and pointing error is derived. A closed-form probability density function (PDF) for pointing errors comprised of boresight and jitter error is derived. Two auto-tracking mechanisms are designed and developed at the transmitter to combat the issues of boresight and jitter error. First, a coarse tracking system is designed with a magnetometer sensor on-board to minimize the boresight error. Any leftover boresight error, known as jitter, is compensated by a fine-tuning and closed-loop feedback system. The proposed auto-tracking mechanism has been experimented with a wide range of non-zero boresight angles, from ±10◦ to ±180◦ with a fine-tuning of jitter up to 8 cm. The performance of the proposed system has been analyzed in terms of outage probability. The analytical results are compared with simulation, measurement, and existing methods. • FSO is considered primarily for establishing point-to-point (PtP) communication. A novel point-to-multipoint (PtM) tracking mechanism for FSO communication is designed and implemented to meet the increasing demand for mobile platforms and mechanisms to establish PtM connections. The alignment mechanism consists of two tracking systems, i.e., coarse tracking and fine-tuning. The coarse tracking is responsible for locating FSO transceivers placed at different positions using magnetometer sensors connected to the cloud. On the other hand, the fine-tuning mechanism overcomes high-frequency amplitude oscillation or jitter with a displacement range of 8 cm. The link is continuously monitored, and if the link is in inoperable condition or not established due to extreme weather conditions, the reference line-of-sight (LoS) angle is updated in the cloud to establish an alternate connection. The stability of the proposed system is analyzed in terms of root locus and step response. The proposed PtM system is validated experimentally with FSO terminals located at different distances over indoor and outdoor environments. The proposed method achieved an alignment time of less than 5 seconds for 10 degrees of misalignment
Tailoring The Magnetoelectric Properties Of Srfe12o19 Via Partial Substitution Of Bi And Mn
No full textThere has been a huge demand nowadays for the magnetoelectric materials for multifunctional devices. The research activity on magnetoelectric multiferroic materials has increased significantly by realizing its potential for meeting the demand for technological need. In this regard, strontium hexaferrite and its derivatives are prepared. In the present thesis, various physical properties (such as structural, surface morphology, electrical, magnetic, etc) along with the magnetoelectric property have been studied for the synthesized compounds. The strontium hexaferrite (SrFe12O19), belongs to the family of M-type hexaferrite. It is a magnetoelectric multiferroic compound having high Curie temperature (TC ~743 K), high electrical resistivity (in the megaohm range at room temperature (RT)), strong net magnetization (20 ..⁄) and a very hard (~ 6 GPa) compound. The SrFe12O19 possesses giant coercivity value (~ 4.5 kOe at room temperature) and high magnetocrystalline anisotropy (~105 erg/cm3). The strontium hexaferrite is extremely flexible to customize its magnetic property or other physical properties in order to meet the desired application requirement. Here, the SrFe12O19 and bismuth (Bi) and manganese (Mn) substituted SrFe12O19 compounds (bismuth (Bi) and manganese (Mn) are substituted in place of strontium (Sr) and iron (Fe) respectively) are prepared using sol-gel auto combustion method. All the prepared samples are confirmed to be phase pure via X-ray diffraction (XRD). The lattice parameters, unit cell volume, lattice strain, etc. are obtained by the refinement of the XRD data via Rietveld technique. The phase purity of the compound SrFe12O19 (SrM), Sr0.98Bi0.02Fe12O19 (SBFO2), SrFe9Mn3O19 (SrM3), SrFe7Mn5O19 (SrM5), and SrFe5Mn7O19 (SrM7) are also confirmed by the neutron diffraction (ND). Both the refinement data confirm the hexagonal crystal structure of the compounds having space group P63/mmc. The bond valence calculation from ND data shows the presence of Fe2+ at 12k crystallographic sites and it indicates an increase of Fe2+ content at 12k-site with increasing Bi in the sample. The distortion calculation from ND suggests an increase in strain at the 2b, 12k-sites due to Bi and Mn substitution. The ND for Mn substituted compounds reveal that Mn ions mostly prefer 2a and 12k- sites. The total magnetic moment (calculated from the respective site occupations) for the Mn substituted compounds, decreases with Mn substitution at room temperature (RT). Similarly, for SBFO2, the total magnetic moment increases as temperature drops. The surface morphology study is carried out via field emission scanning electron microscopy (FESEM) and all samples grain size are in the range ~0.5 μm - 1 μm order. The FESEM micrographs show grains-shape transformation from hexagonal to rhombohedral for SBFO2 and SrM7. The highly strained compound Sr0.99Bi0.01Fe12O19 (SBFO1) and SrM5 grains are most irregular-shaped grains as well as irregular sizes and grain arrangements are also different from rest of the compounds. The strain created in the systems is also observed in the Raman spectra, where Raman spectra changes drastically for Mn substituted compounds and in case of SBFO1, a new Raman mode evolves. The X-ray photoelectron spectroscopy (XPS) measurement is also performed for selected compounds and it confirms the existence of desired elements in the respective compounds. Respective element’s valence states are also deduced from high-resolution XPS spectrum. The XPS data stands by the bond valence calculations. The iron ion exists in both 2+ and 3+ states in SBFO2, SrM5, and SrM7. Similarly, the Mn exists in different states (2+, 3+, and 4+) in SrM5 and SrM7. The Mn3+, Mn4+ increases and Fe3+ content decreases with increasing Mn content. The magnetization variation as a function of temperature (MT-curve) for all the compounds show irreversibility in zero field cooled (ZFC) and field cooled (FC) data right below the Curie temperature (TC) and the Hopkinson peak is observed in the ZFC data. The TC varies in a short temperature range (733 K-739 K) in the Bi substituted compounds whereas in the case of Mn substituted compounds, the TC drops drastically (~620 K for SrM3, ~517 K for SrM5, and ~429 K for SrM7). The TC shifts to higher temperature under the application of magnetic field. The temperature dependence of saturation magnetization of all the compounds follows the Bloch relation and the exponent obtained is close to the theoretical value (1.5). All compound’s saturation magnetization decreases with substitution except for SBFO1, where slight increase is observed. However, the coercive field shows an unusual behaviour with temperature and this has been fitted using an equation devised by taking into consideration of all the three anisotropies (i.e. magnetocrystalline anisotropy, shape anisotropy, and stress anisotropy). The critical exponents are calculated for SrM, SBFO1, SBFO2, SrM3, and SrM5 compounds at the ferromagnetic-paramagnetic phase transition (second-order phase transition) boundary by using modified Arrott plots. These critical exponent values are slightly overvalued as per mean-field theory. Interestingly, for the Mn substituted SrM5 and SrM7, the field varied magnetization data (MH loop) has contribution of two magnetic phases (FiM1 and FiM2) and it is confirmed through the Derivative - Deconvolution – Selective Integration (D-D-SI) method. The presence of the second magnetic phase is also seen in the ZFC curve of SrM5. From SrM5 onwards, the growth of another magnetic phase (FiM2) of lower coercivity apart from the parent phase (FiM1) of higher coercivity is seen. The FiM2 phase is found to be increasing with the Mn content in the sample (16% for SrM5 but 66% for SrM7). Although the magnetization for both FiM1 and FiM2 decreases with increasing temperature, for SrM7 the FiM2 phase is found to persist till higher temperatures compared to the FiM1 phase, whereas just the opposite behavior is observed for SrM5 sample. It thus suggests a transformation of compound from high magnetic anisotropy (SrM) to low magnetic anisotropy (SrM7). The FiM2 phase 2() and 2() decreases with rising temperature. The presence of Fe2+ at the 12k site and the rise of Fe2+ percentage by Bi are supported by the Mössbauer data of SrM, SBFO1, and SBFO2. The impedance and modulus study of parent and the Bi substituted compounds suggest the presence of two relaxation mechanisms and the two contributions are coming due to grain and grain boundary. A transition in the relaxation mechanism appears from grain dominated (below 150 K) to grain boundary dominated (above 150 K). In the case of the Mn substituted, the presence of only one relaxation mechanism is dominating (i.e. grain). The two anomalies that appeared in the output of impedance and modulus for the parent and Bi substituted compounds are coming from the magnetic blocking temperature i.e. 75 K and another one due to the grain-grain boundary relaxation transition. The extracted resistance value (of grain and grain boundary) for SrM, SBFO1, and SBFO2 decreases with temperature. The grain related capacitance for SBFO1 is found to be ∼50 nF, and this value is ∼10 times higher than SrM, SBFO2, whereas the grain resistance has dropped by 10 times for SBFO1. Similarly, grain resistance obtained from impedance shows that Mn substitution decreases the resistance by two order as compared to SrM. The SrM, SBFO1, and SBFO2 resistivity data follow the nearest neighboring hopping model, but the resistivity data of Mn substituted compound confirms the conduction via variable range hopping and nearest neighboring hopping. The magnetoresistance measured at various sub-room temperatures for Bi substituted compounds (i.e. SrM, SBFO1, and SBFO2) shows the interplay of anisotropy magnetoresistance (AMR) and giant magnetoresistance (GMR). Low temperature data are dominated by GMR and gradual participation of AMR increases as room temperature is approached. The linear magnetoelectric coefficient αd⁄ (in mV∙cm−1∙Oe−1) for SrM is found to be 0.33(2) at 125 K, and this value decreases gradually to 0.27(1) at 300 K. The SBFO1 sample displayed the highest (even 10% higher than the SrM sample) value of αd⁄ at low temperature. Unfortunately, the increased value of αd⁄ is also accompanied by a drastic reduction in its magnitude for temperatures higher than 200 K, due to the increased electrical conduction which in SBFO1 is ∼94% higher than the parent. Similarly, in the Mn substituted compound, SrM5 compound shows maximum ‘αd⁄’ value, 0.83(2) mV∙cm−1∙Oe−1 and this ‘αd⁄’ value is ~2.5 times higher than that of the parent sample at low temperature. This enhancement is due to the strain produced in the SrM5 system. The influence of strain is also seen in the quadratic magnetoelectric coefficient (2βd⁄). The ‘2βd⁄’ value is also highest for SrM5 (i.e. 2.83(2) mV∙cm−1∙Oe−2) as compared to the other compounds
Analysis of Intrusion Data using Scalable Machine Learning Techniques
No full textThis thesis presents our work concerning the design and implements intrusion detection systems using scalable machine learning approaches. To design IDS, a large amount of threat signature is required for the machine learning-based detection approach. Hence our first work focuses on towards preparation of a new intrusion dataset with the latest threat signature. A testbed is created in our lab to launch the attacks, capture the attack pattern, and store them in packet captured (PCAP) data format. The most crucial and tedious work is feature generation, preprocess of features, assigning the class label, and making it compatible for data analysis. To implement the detection engine, we start with signature-based detection approaches using Snort and BroIDS. The rules were written for many attacks by analyzing their signatures and tested them in our laboratory. This approach is more suitable for known attacks. This approach’s main disadvantage is the impossibility of detecting new intrusions because they only look for patterns that match the rules stored in the databases. The database needs frequent updates as recent attacks are discovered. Human intervention is required for attack detection, signature generation, rule generation, and distribution. The anomaly-based approach is implemented using unsupervised and supervised techniques to overcome the disadvantages of signature-based detection approaches. The unsupervised anomaly detection is having a good idea of taking the unlabeled data and based on characteristics of data distribution, the class label is predicted for each object. Because of this property, there is no need for labeled training data. This method is suitable for finding new attacks without prior knowledge about the attack. However, it raises false alarms due to malicious activity that does not significantly change from regular activity in some cases. Further, to improve the detection process, the ensemble approach comes into the picture and takes a vital role in constructing a robust detection engine and achieve better results as compared to earlier techniques. It divides the problem with all the learners who participated in the training process and combines them. The data analysts use the ensemble approach using popular machine learning, design pattern recognition, data mining, neural networks, and measurement techniques to solving various problems. It is well known that an ensemble is typically similar to an individual learner, and ensemble strategies have made impressive progress in the different real-world tasks. The ensemble approach is based a supervised as well as unsupervised learning algorithms. In the ensemble method, the number of weak learners is trained, compares the learning error, and rearranges the data distribution by multiplying the weight vector until the error is minimized. Making predictive models from the available knowledge base is known as the learning or training process. The trained model is known as a hypothesis. The ensemble approach aims to combine multiple hypotheses to form a better hypothesis that provides a stable output in terms of accuracy and the least error using the same weak learner. The ensemble approach can be more flexible by adopting the model optimization techniques. But it increased the complexity of the model and needs high computational resources to process the data. Big Data Environment makes the existing developed algorithms scalable that support a high volume of data for analysis. As per the study, the bottleneck during processing does not occur in distributed processing. As per the Map-Reduce functions, the data calls the program for processing. The probability of bottleneck can be avoided in the large-scale analytic process due to the program’s memory size being negligible compared to the size of data. The significant advantage is that the program comes to the data for processing. In conventional data processing, the program calls the data as an argument that leads towards bottleneck and requires high computational resources during execution. Big Data Processing is quite different. The size of the program is significantly less as compared to the data. The Map-Reduce, Spark, and HDFS (Hadoop Distributed File System) provide reliability, flexibility, high performance, and efficiency in storing, managing, monitoring, processing, and visualizing the data. An efficient IDS approach is implemented to deal with the imbalanced data using supervised and unsupervised techniques on a scalable platform to detect minority attack classes. For more sophisticated attacks and to address multi-class problems, a particular type of threat, a deep neural network-based detection approach is implemented. Deep Learning Models require a large amount of training data during the learning process. The intrusion dataset and our prepared dataset are used to train, validate, and test the models. Empirically, we have observed that the DNN models able to detect the broad attack classes more efficiently
Refinement and Processing of Steel Microstructure Images Facilitating Automated Heat Treatment Process Prediction
No full textSteels are alloys of iron and carbon (with up to 2.1 weight percent), widely used in construction, electronics, automotive and transport, energy, packaging, containers, appliances, etc. The variety of properties that the different grades of steel (used in these different industries) possess, is high. Different applications and various industries require different physical, mechanical and chemical characteristic features of steels, which necessitates making of different grades of steel. The application specific properties or features of steels are obtained by subjecting the metallic alloy to thermomechanical, or heat treatment procedures. Majority of these heat treatment procedures in the steel industry generally involve a lot of human resources as automation in this sector still requires more advancement. Also, these processes are highly resource intensive and time consuming. If because of these manual errors, the desired grade/quality is not achieved in the final product, it leads to a lot of wastage of resources. To curb this, automated or computerized simulations of these heat treatment procedures are increasingly being considered as an alternative. In this thesis the challenge of digitizing plain carbon steel microstructure images is addressed, which would aid in the automation of steel heat treatment processes in the metallurgy industry. This kind of automation will help in quicker novel material discovery and facilitate easier inverse design of materials, i.e., designing materials given a set of desired properties. Multiple neural network architectures for the refinement (denoising and segmentation) of raw steel microstructure images are developed, suitable for the above purpose. During the process of building optimal deep learning architectures, it was found that generative models are better suited for handling the variable types and amounts of noise found in plain carbon steel microstructure images. Also, using generative models for microstructure image refinement tasks like denoising and segmentation, helps us bypass deploying very deep convolutional neural network models, which have high computational demands, and take more time to converge/train successfully. Finally, in this thesis, a machine learning framework is elucidated, for the accurate prediction of a suitable thermomechanical treatment process, in order to attain a target grade of steel from a given initial steel sample. Conventional image featurebased or texturebased tools, when combined with machine learning techniques prove to be very useful in microstructure classification/prediction. Deep learning based classifiers have also been used, and the conventional machine learning techniques report comparable performance with these stateoftheart deep learning models
Compressed Multimedia Forgery Detection through Blind Digital Forensics
No full textIn today’s cyber world, digital images and videos act as the most frequently transmitted information carriers. However, the present day easy availability of low–cost image and video processing software and desktop tools, having immense number of multimedia manipulating features, pose threat to the fidelity of digital multimedia data. Digital images and videos being the major sources of evidence towards faithfulness of any event for law enforcement across the world, as well as in media and broadcast industries, maintenance of their trustworthiness and reliability is a major challenge in today’s digital world. This research deals with compressed image and video forgery detection. Almost all present–day digital devices use lossy/lossless compression formats for data storage. This research aims to exploit the effects of lossy/lossless compressions in multimedia files for detection of forgery. Joint Photographic Experts Group (JPEG) re-compression based image forensics has been a largely investigated area in the forensic community, over the recent years. This mainly involves investigating different compression quality factors in an image, by exploiting artifacts left behind by the compression stages. This thesis presents a deep Convolutional Neural Network model, that detects both double as well as triple compression in JPEG images. This thesis also addresses the major present-day challenges associated with video forgery detection, aimed towards solving the practical problems encountered by a forensic analysts during video tampering detection. Video forgery can be broadly classified into inter-frame and intra-frame forgery types based on their operational principles. This research focuses on the development and performance analysis of inter- as well as intra-frame video forgery detection and localization techniques, while addressing the certain additional challenges i.e., scenarios where number of forged frames is integral multiple of Group of Pictures (GOP), video compressed with dynamic GOP etc. This thesis presents statistical as well as deep learning based inter-frame forgery detection techniques. This thesis also investigates and develops a capsule network based forensic technique for detection of object based intra-frame forgery in surveillance videos, while improving the performance for low bit rate compressed videos
Performance Enhancement of Vehicular Ad hoc Network (VANET) for Real-life Applications in Smart Cities
No full textVehicular ad hoc network is one of the fastest-growing research areas within a decade in industry and academia. The performance of VANET is dependent on the communication between vehicles and RSUs. Nowadays, vehicles are not considered a traditional carrier. Still, the technological advancement in the electro-mechanical-humanity department, availability of various sensors at affordable cost, a communication unit, availability of vehicular standard for effective communication, availability of 5G for the quick transmission of data make the vehicle smart and intelligent. It is used for a wide range of autonomous applications. The demand for various autonomous applications in smart cities and the wide range of safety and non-safety applications of VANET motivates to enhance the performance of VANET for the real-life application of the smart cities. Selection of optimum path from source to destination, the quick transmission of data from the source and quick reception of data at the destination for low density of vehicles, the quick transmission of information for dynamic destination, and detection of faulty OBU and isolate the faulty vehicle from routing for the successful implementation of the application through VANET are the objectives of the thesis. Simulation through SUMO and OMNET++ and the testbed experiments are performed for the performance measurement of the proposed work. A position based routing protocol is proposed for the transmission of emergency messages. During message communication, the vehicles communicate with the neighboring vehicle in the same direction and take advantage of vehicles moving in the opposite directions in another lane. Also, the routing protocol considers the combination of parameters such as the shortest path, the density of vehicles, and the delay. To make the network scenario realistic, it assumes an important network impairment such as attenuation. The simulation result shows that the proposed work performs better than the existing protocol in terms of E2E delay, number of hops, number of vehicular gaps, and Total Service Time (TST). NIB based routing is proposed for the health parameters monitoring at the health observation center. During COVID-19, the transmission of health parameters from the quarantine center to the health observation center and analysis of the health parameters at the observation center through VANET is proposed. The transmission of the health parameters is presented using an automated routing through NIB, vehicles, and RSU. Analysis and classification of health parameters at the observation center are done using one against one (OAO) method of support vector machine. The performance of the proposed routing is compared with P-GEDIR, A_STAR, and GSR, and performs better in terms of end-to-end viii delay, number of network gap, packet delivery ratio, etc. Health parameters analysis is done using the parameters precision, recall, f1-score, and support. The testing accuracy for the health data analysis phase is determined as 96.8%. Location service based routing is proposed for location identification which is suitable for the mobile destination. The house location identification is considered based on the house owner’s name or house number, including the patient’s house for this purpose. House location information is stored in a distributed open-source document database called MongoDB in which the location of the house is searched either in terms of a house number or house owner name. The searching of the house location information is done using a tool node-red. The performance of the proposed work is compared with the location service GLS and HLS and performs better in terms of query success rate, location MAC bandwidth consumption, and request travel time. A fault based routing algorithm considering a complete fault set is proposed to diagnose the hard fault, three categories of the soft fault, i.e., permanent, intermittent, and transient, for the faulty OBU of the VANET. Along with soft fault detection, soft fault classification is done. The information regarding the faulty vehicle is transmitted from the RSU to the faulty vehicle with the help of an automated routing. The K-S test does the hard fault detection with a timeout mechanism. The soft fault detection is done by the chi-square test. Both the hard fault and the soft fault are evaluated by the parameters FDA and FAR. Soft fault classification is done using the ECOC method of the SVM multi-class classifier. The evaluation of the proposed routing method is carried out by E2E delay, the number of hops, and network gaps and performs better as compared to GSR, A_STAR, and Bhoi et al
Correlating The Structure of Hydrogels to Their Diffusion Dynamics & Design and Construction of Fluorescence Correlation Spectroscopy Setup
No full textThe fundamental phenomenon responsible for the movement of agents in the body such as drug molecules, signaling molecules and essential nutrients is diffusion. The transport of the material by diffusion is of great importance in several biological and industrial applications. In applied sciences such as biochemical engineering and pharmaceutical industries, diffusion plays a crucial role to predict diffusant transport ex: drug release rates. Scientists in the field of physical chemistry, on the other hand often study diffusion to extract information on transport properties. The study of diffusion in real-life systems is often complex as many factors govern the diffusion phenomenon such as the inherent structural organizations or the presence of other molecules. Apart from the hindrance experienced due to all these factors, interactions also influence the diffusion process. All these environmental constraints present in the system makes it extremely difficult to figure out the exact parameters affecting the diffusion process. To circumvent this problem, “model systems” have been used in the past as an alternative to real-life systems to study the diffusion process. The main advantage of these model systems is the freedom to control and tune individual parameters and study the effects on the diffusion process. The studies on the model systems can be designed to vary the amount of crowding, interactions, or the presence of other molecules. This will help us to understand the diffusion processes which can then aid in deriving results for real-life systems. Among the various model systems that have been considered, hydrogels have been extensively used to study and understand the diffusion mechanism. Because of its unique mechanical properties and high water content, hydrogel resembles natural tissue more than any other material. Hydrogel is considered as an excellent vehicle for drug delivery and other applications due to its capability to constrain the diffusive movement of the solute molecule. The diffusion of solutes in hydrogels has found applications in a wide range of fields such as chromatography, prosthetic applications, cell encapsulation for fermentation and biomedical application among others. One of the significant parameters that affect the diffusion mechanism in hydrogels is the hydrogel network architecture. The hydrogel structure can be tuned by varying many parameters like polymer and crosslinker concentration, crosslinker types, pH, solvent, temperature, the addition of external moieties such as surfactants etc. It is therefore imperative to have a comprehending knowledge of the parameters governing solute diffusion in hydrogels as well as how the hydrogel network architecture affects the diffusion process. Taking into consideration all the above facts the research works presented in the current thesis have been framed under two main objectives. The first objective is to investigate how the diffusion dynamics are influenced by the variation in various parameters such as polymer concentration, crosslinker concentration, charge density etc. The second objective xxii is to try and correlate the observed diffusion dynamics to the change in the hydrogel network structure. The effect of the addition of external moieties such as surfactants of different head groups and concentrations on the structure and dynamics of gellan gum hydrogels were investigated. A strong interaction is seen between gellan gum and oppositely charged cationic surfactant, hexadecyltrimethylammonium bromide (CTAB) whereas rather weak or minimal interactions are observed when either anionic surfactant, sodium dodecyl sulfate (SDS), or non-ionic surfactant, Triton X-100 is added to the system. An interesting cross-over from stretched to compressed exponential was seen when CTAB was added beyond critical micellar concentration to the system, which was not evidenced for the other two surfactants. The attachment of CTAB clusters or micelles can lead to the distortion of the hydrogel network which then acts as stress releasing points resulting in the observed crossover. These results may have important repercussions for the use of polymer surfactant systems as potential products. The effects of mixed salts on the structure and dynamics of the physical gel of gellan gum were evaluated. It was seen that when mixed salts were present in the system, the storage modulus of the hydrogels decreased as compared to the pure hydrogels. The main reason for this type of decrease in the storage modulus as well as the relaxation time may be that the addition of mixed salts may not reduce the electrostatic repulsion between the gellan gum helices more effectively than the monovalent or the divalent salt alone. Thus there might be a hindrance to the formation of the strong networks, which leads to a decrease in the storage modulus values when two types of salts are present. The effect of monomer concentration and charge density on the structure and dynamics of a chemical hydrogel of acrylamide/sodium acrylate was investigated. With increasing total monomer concentration, the storage modulus of the hydrogel increased. However, when the charge density was varied, the storage modulus of the hydrogels first increased and then decreased monotonically. In the presence of Bisacrylamide crosslinker an interesting change in dynamics was seen which was dependent on both the monomer concentration as well the charge density of the system. The hydrogel changed from neutral polymer to charged polymer after the charges were introduced to the system. Owing to this charged nature an extra mode is observed in the relaxation dynamics. Moreover, the degree of spatial gel inhomogeneity was seen to decrease moderately with an increase in sodium acrylate concentration and decreased significantly with a decrease in total monomer concentration. The effects of varying the types of crosslinkers on the triple dynamics were investigated further. For this purpose, two multifunctional crosslinkers (divinyl sulfone, DVS and glutaraldehyde, GLU) were used. Firstly, the effect of varying the total monomer xxiii concentration was studied by rheology and it was seen that in both crosslinkers the storage modulus increased with increasing total monomer concentration. Also, it was evidenced that the DVS crosslinker produced the hydrogels with higher storage modulus as compared to the GLU crosslinker at all the total monomer concentration range studied. From dynamic light scattering measurements we evidenced an extra mode in both the crosslinker systems and for all the total monomer concentrations studied. It was seen that this extra mode was observed for all the hydrogels at and after 80% SA concentration. Neither the monomer concentration nor the type of crosslinker used had any effect on the emergence of the new relaxation mode as seen earlier. There was no particular trend in the onset of triple mode which was observed before while using the Bisacrylamide crosslinker. The observed difference may be due to the different crosslinking mechanisms of the crosslinkers studied. As part of the research work, we have designed and constructed a Fluorescence correlation spectroscopy (FCS) instrument in-house. This was used to probe the structure and dynamics of hydrogels under varied conditions. We strongly believe that these studies will support the fundamental understanding of real-life systems and their operation in complex environments. These studies can be utilized to design the structure of hydrogels that can be used for specific applications such as targeted drug delivery, although in a broader sense can be applied to a myriad of applications
Numerical Investigation of Power-law Fluid Flow and Heat Transfer over a Rotating Elliptic Cylinder in Laminar Flow Regime
No full textA two-dimensional laminar flow of non-Newtonian fluid over a heated rotating elliptic cylinder is numerically studied in the present work. The flow field around the cylinder is computed by solving non-linear partial differential governing equations using ANSYS FLUENT (2015). The computational domain and mesh are generated and implemented using ANSYS Workbench. The sliding mesh method (SMM) is implemented to capture the transient flow behaviour around the rotating elliptic cylinder. The non-Newtonian shear-thinning and shear-thickening fluid behaviour is modelled by power-law model. The aim of the work is to study the effect of cylinder rotation, fluid behaviour and cylinder’s aspect ratio along with Reynolds and Prandtl number on fluid flow and heat transfer phenomena. The range of dimensionless parameters considered are aspect ratio of the cylinder = 0.1, 0.4, 0.7 and 1.0, power-law index 0.4 ≤ ≤ 1.8, rotational speed 0.5 ≤ ≤ 2.0, Reynolds number 5 ≤ ≤ 100 and Prandtl number 1 ≤ ≤ 100. The engineering parameters, like drag and lift coefficient ( and ), stream function and vorticity contours, local and average Nusselt number () are reported to demonstrate the effect. Streamline and vorticity patterns in the vicinity of the cylinder illustrate the wake formation and vortex shedding phenomenon in the rear part of the cylinder. Due to the cylinder rotation, vortices attached to it either merge and/or detach to/from the cylinder, and are termed as enveloping and detached vortex, respectively. Their size and characteristics largely depend upon , , , and . The power-law index (), aspect ratio (), and rotational speed () encourage the formation of enveloping vortex; however, they have a negative impact on detached vortex. The average values of the drag coefficient are seen to decrease with and . And, a non-monotonous trend is observed with the power-law index () and aspect ratio (). The negative lift coefficient is observed due to the anticlockwise rotation of the cylinder for all parameters considered here. At a higher Reynolds number, = 100, the flow around the rotating elliptic cylinder exhibit the Von-Karman vortex shedding phenomenon. However, the high rotation of the cylinder ( > 1.0) obstructs the shedding of the vortices significantly. For shear-thinning and Newtonian fluids, the vortex at the top edge starts hovering in the vicinity of the cylinder. The vortex is named as Hovering vortex (HV). The shearthinning behaviour of the fluid encourages the formation of HV; however, HV is not observed for shear-thickening fluids. The HV causes a long shedding period. The variation of the local Nusselt number on the surface of the cylinder is seen to be maximum at the edges of the cylinder for all , , , and considered here. As expected, the Prandtl number () and shear-thinning behaviour ( < 1.0) encourage the heat transfer from the cylinder. In addition, the Nusselt number is more at higher rotational speed of the cylinder. A correlation is presented for the time average surface Nusselt number () as the function of Prandtl number (), fluid behaviour (), and rotational speed (), i.e. = (, , , and ) to find the Nu at intermediate values of , , , and