Browsing by Author "Viswanathan, Sangeetha"

Now showing 1 - 20 of 26

Anomaly-based Intrusion Detection using GAN for Industrial Control Systems
(IEEE, 2022) Viswanathan, Sangeetha
In recent years, cyber-attacks on modern industrial control systems (ICS) have become more common and it acts as a victim to various kind of attackers. The percentage of attacked ICS computers in the world in 2021 is 39.6%. To identify the anomaly in a large database system is a challenging task. Deep-learning model provides better solutions for handling the huge dataset with good accuracy. On the other hand, real time datasets are highly imbalanced with their sample proportions. In this research, GAN based model, a supervised learning method which generates new fake samples that is similar to real samples has been proposed. GAN based adversarial training would address the class imbalance problem in real time datasets. Adversarial samples are combined with legitimate samples and shuffled via proper proportion and given as input to the classifiers. The generated data samples along with the original ones are classified using various machine learning classifiers and their performances have been evaluated. Gradient boosting was found to classify with 98% accuracy when compared to other
Assessment of an ant-inspired algorithm for path planning
(Elsevier, 2022) Viswanathan, Sangeetha
The demand for path planners for a variety of applications has significantly increased over the past decade. The correct choice of a distance metric will be of utmost importance for an efficient path planner. The underlying connectivity of the roadmaps produced by the planner are determined by the metrics. A study was conducted in this chapter for the proper choice of planner metrics. Five metrics from the literature were chosen and implemented in a gain-based ant colony optimization (GACO) algorithm. Results are analyzed against parameters, such as time taken, length of the path, and turn characteristics. Finally, the GACO with the chosen metric was implemented using different satellite images from the International Society for Photogrammetry and Remote Sensing and compared against existing algorithms with respect to performance.
Assessment of cloud vendors using interval-valued probabilistic linguistic information and unknown weights
(Wiley, 2021-05) Viswanathan, Sangeetha
Cloud vendors (CVs) play an indispensable role in the development of IT sectors and industry 4.0. Many CVs evolve every day, and a systematic selection of these is becoming substantial for organizations. Literature studies have shown that multicriteria decision-making (MCDM) is a powerful tool for systematic selection. However, the major issue with the state-of-the-art models is that they do not effectively represent uncertainty. Moreover, the personalized selection of CVs based on user queries is not prominent in an MCDM context. In this paper, to circumvent these issues, a new decision framework is proposed that utilizes a generalized preference style called interval-valued probabilistic linguistic term set (IVPLTS). This preference style considers occurring probability values as interval numbers instead of a single precise value, which provides flexibility during preference elicitation. Initially, missing values are imputed systematically by using a case-based method. Then, the consistency of these preferences is checked using Cronbach's alpha coefficient, and the inconsistent preferences are repaired rationally by using an iterative method. A programming model is proposed for determining the weights of the evaluation criteria. Furthermore, Maclaurin symmetric mean (MSM) is extended to IVPLTS for aggregating preferences from each expert. The interval-valued probabilistic linguistic comprehensive (IVPLC) method is proposed for prioritizing CVs in a personalized manner. Finally, the framework's practicality is validated by using a case study of CV selection for an academic institution; strengths and weaknesses of the framework are conferred by comparison with extant CV selection models.
Classifying DNS over HTTPS Malicious/Benign Traffic Using Deep Learning Models
(IEEE, 2023) Viswanathan, Sangeetha
As we live in an era where privacy over the Internet has become rudimentary, protocols like DNS over HTTPS (DoH) and DNS over TLS (DoT), which promote encryption, have become popular. While these protocols were introduced to overcome the drawbacks of DNS protocol, even DoH has some security issues that need to be tackled to prevent any misuse. Herein, we implemented deep learning models to classify DNS over HTTPS traffic and found the most efficient method in regard to time-required complexity and computational requirements. Previous studies have used a variety of features from datasets to identify malicious activities. Although machine learning and deep learning models are commonly used, they require more human intervention. These models are also more computationally complex, as one is required to tune the model and its parameters for accurate results. In comparison, some deep learning models are more efficient as they work well without any human intervention and are capable of parameter tuning by themselves. In this work, we used the CIRA-CIC-DoHBrw-2020 dataset and performed data imbalance handling, one hot encoding, and feature selection to create a model that can be used for a more generalized environment. We implemented long short-term memory (LSTM), bidirectional LSTM (BiLSTM), and gated recurrent unit (GRU) models to classify DoH traffic with high accuracy. Although the mentioned models produced good accuracy, the BiLSTM model performs better than the LSTM model in the time taken for prediction and accuracy; the GRU model outperformed both LSTM and BiLSTM models in terms of accuracy, computation time, and computation complexity. Hence, it is more efficient than both LSTM and BiLSTM models.
Cluster analysis of breast cancer data using Genetic Algorithm and Spiking Neural Networks
(IEEE, 2015) Viswanathan, Sangeetha
Breast cancer is taking a large toll in the present scenario. Many computer aided diagnosis are been developed to detect breast cancer. The detected breast cancer is also classified according to their subtypes. In the absence of a class definition, analyzing the cancer types is huge some task. Clustering the breast cancer data is a process that merges the feature selection process and the process of defining the class labels for the data. The proposed work has four stages which include preprocessing, feature selection, feature clustering and cluster validation. This paper uses a Spiking Neural Network that is been trained with an Evolution topology algorithm and Genetic Algorithm is used to select the features from the dataset. The result of the network will cluster that classifies the data into abrupt types. The clusters are then validated using DB index
Cluster Analysis of Breast Cancer Data Using Modified BP-RBFN
(CRC Press, 2020) Viswanathan, Sangeetha
According to the American Cancer Society, around 17,62,450 women have breast cancer in the United States. Though there are many computer-aided diagnosis systems for detecting cancer, the chances of survival of the patients are essential for an efficient cancer management system. This chapter proposes a model for analysing the cancer data using a modified back propagation-based radial basis function neural network. The proposed clustering using BPRBFN has four modules: preprocessing, feature selection, feature clustering and cluster validation. Feature selection is performed using a genetic algorithm. Out of 32 features, 10 essential features are selected and proceeded for further clustering. Radial basis function neural network that is learned using a backpropagation algorithm is used for clustering because of its best approximation. The network is trained using six different training algorithms and compared to find the best training algorithm for an optimal clustering. The final results are validated using the mean-square error index and regression fit value. Wisconsin Breast Cancer (Diagnostic) Dataset is used for all simulations throughout the work.
Energy-efficient green ant colony optimization for path planning in dynamic 3D environments
(Springer, 2021-01) Viswanathan, Sangeetha
With the proliferation in demand for navigation systems for reconnaissance, surveillance, and other day-to-day activities, the development of efficient and robust path planning algorithm is an open challenge. The uncertain and dynamic nature of the real-time scenario imposes a challenge for the autonomous systems to navigate in the environment, avoiding collision with the moving obstacles without compromising on the energy-time trade-off. Motivated by this challenge, an efficient gain-based dynamic green ant colony optimization (GDGACO) metaheuristic has been proposed in this paper. The energy consumption while path planning in a dynamic scenario will be humongous owing to its nature. The proposed algorithm reduces the total energy consumed during path planning through an efficient gain function-based pheromone enhancement mechanism. The memory efficiency of Octrees is incorporated for workspace representation because of its ability to map large 3D environments to limited memory. Comprehensive simulation experiments are conducted to demonstrate the efficacy of GDGACO. Results are analysed through comparison with other methods in terms of path length, computation time, and energy consumed. Also, the results are verified for statistical significance.
Evaluation of Distance Functions for the Comparison of Gradient Orientation Histograms
(Indian Journal of Science and Technology, 2015) Viswanathan, Sangeetha
Local features of an image are used in many computer vision applications such as object detection and scene matching. The gradient orientation histogram is used by many local features such as Scale Invariant Feature Transform (SIFT), a widely used image local feature. This paper discusses various distance functions that can be used to measure the similarity between the local features described by the gradient orientation histogram. A distance function, based on the quadratic form is proposed for the SIFT descriptor. The state of the art distance functions - Euclidean, Chi-square, Manhattan and the proposed quadratic form based distance function are calculated between the features extracted from the images. Nearest neighborhood ratio strategy is used to find the corresponding features based on the distance measure. Correct matches are estimated using the ground truth transformation function between the images, present in the form of homograph matrix. It is experimentally found that the proposed distance function has an execution time reduced by 21% compared to the Euclidean distance for a similar accuracy performance. The proposed distance retrieves more number of correct matches compared to the modified Earth Mover distance which is fastest among the evaluated distance functions. The future work will be aimed at improving the time taken for computing the distance matrix between the feature sets and a better strategy for computing the matches.
A Fuzzy Gain-Based Dynamic Ant Colony Optimization for Path Planning in Dynamic Environments
(MDPI, 2021-01) Viswanathan, Sangeetha
Path planning can be perceived as a combination of searching and executing the optimal path between the start and destination locations. Deliberative planning capabilities are essential for the motion of autonomous unmanned vehicles in real-world scenarios. There is a challenge in handling the uncertainty concerning the obstacles in a dynamic scenario, thus requiring an intelligent, robust algorithm, with the minimum computational overhead. In this work, a fuzzy gain-based dynamic ant colony optimization (FGDACO) for dynamic path planning is proposed to effectively plan collision-free and smooth paths, with feasible path length and the minimum time. The ant colony system’s pheromone update mechanism was enhanced with a sigmoid gain function for effective exploitation during path planning. Collision avoidance was achieved through the proposed fuzzy logic control. The results were validated using occupancy grids of variable size, and the results were compared against existing methods concerning performance metrics, namely, time and length. The consistency of the algorithm was also analyzed, and the results were statistically verified.
GAN-Based Anomaly Intrusion Detection for Industrial Controller System
(Springer, 2023) Viswanathan, Sangeetha
Industrial controller system (ICS) is becoming more and more important in daily lives. In recent years, ICS has become more frequent targets of cyberattacks. In addition to the system, the environment is also significantly impacted by the ICS cyberattack. The main aim of ICS intrusion detection is a process of anomaly detection because cyberthreats cause anomalies to occur in the ICS and components under its control. With a machine learning or deep learning aid, the IDS can produce precise detection outcomes. Though machine learning models can be used to detect cyberattacks, there is a challenge in handling imbalanced real-time data. In this paper, we have implemented generative adversarial networks, to resolve the issue of imbalance in datasets by creating class-specific adversarial samples and further detecting anomalies with greater efficiency. This proposed method is tested on Secure Water Treatment Dataset.
A Hybrid Gain-Ant Colony Algorithm for Green Vehicle Routing Problem
(IEEE, 2022) Viswanathan, Sangeetha
Increasing carbon emissions, and thus footprint, is one of the main reasons for the imbalance in environmental sustainability, which is primarily contributed to transportation. Transportation is a core functionality of logistics distribution and supply chain. In this paper, a hybrid gain-ant colony optimization and fruit fly optimization algorithm for green vehicle routing problem is proposed to plan shortest paths with reduced total fuel consumption efficiently. The proposed algorithm was simulated using the Erdogan and Miller Hooks dataset and compared with best-known solutions and existing methods.
An Intelligent Gain based Ant Colony Optimisation Method for Path Planning of Unmanned Ground Vehicles
(DRDO, 2019) Viswanathan, Sangeetha
In many of the military applications, path planning is one of the crucial decision-making strategies in an unmanned autonomous system. Many intelligent approaches to pathfinding and generation have been derived in the past decade. Energy reduction (cost and time) during pathfinding is a herculean task. Optimal path planning not only means the shortest path but also finding one in the minimised cost and time. In this paper, an intelligent gain based ant colony optimisation and gain based green-ant (GG-Ant) have been proposed with an efficient path and least computation time than the recent state-of-the-art intelligent techniques. Simulation has been done under different conditions and results outperform the existing ant colony optimisation (ACO) and green-ant techniques with respect to the computation time and path length
A Modified Ant Colony Optimisation based Optimal Path Finding on a Thematic Map
(IEEE, 2019) Viswanathan, Sangeetha
Increasing demands for unmanned systems and the availability of high resolution satellite images have been promoting researchers to contribute innovations to increase the robustness and efficiency of the optimal path planning. An effectively classified satellite image and a robust path planning strategy are highly desirable in finding an optimal path. In this paper, satellite images from ISPRS are classified to identify the traversable areas using a Deep Convolutional Encoder-Decoder architecture-Seg Net and cost map is generated. Using the cost map, Modified Gain based Ant Colony optimization(MGACO) is introduced to find an energy efficient path. The path is finally smoothened using Bezier Spline approximation. MGACO has been compared with up-to-date algorithms and results outperform existing methods in terms of run time and length of the path.
A Modified Fuzzy A* Based Inference System for Path Planning in an Unknown Environment
(IEEE, 2018-05) Viswanathan, Sangeetha
One of the major challenges in a path planner system is finding the shortest path in the least time and cost. In this paper, fuzzy logic based A* algorithm is discussed. A* search is a heuristic method and when it is combined with Fuzzy logic, will yield greater global convergence. The proposed method addresses the path -searching behavior with minimum time and risk. In the fuzzy logic architecture, diagonal front-right distance, turn, goal distance, angle and speed are given as the inputs to the fuzzy inference system. Mamdani systems is used as Inference system for a better simulation the experiment is carried out in MATLAB environment and it is found that Fuzzy A* shows a shorter path in less time than the existing A* method.
A Normalized Rank Based A* Algorithm for Region Based Path Planning on an Image
(Springer, 2019-04) Viswanathan, Sangeetha
With the development of many autonomous systems, the need for efficient and robust path planners are increasing every day. Inspired by the intelligence of the heuristic, a normalized rank-based A* algorithm has been proposed in this paper to find the optimal path between a start and destination point on a classified image. The input image is classified and a normalized rank value based on the priority of traversal on each class is associated with each point on the image. Using the modified A* algorithm, the final optimal path is obtained. The obtained results are compared with the traditional method and results are found to be far better than existing method.
Novel Fuzzy Clustering Methods for Test Case Prioritization in Software Projects
(MDPI, 2019-11) Viswanathan, Sangeetha
Systematic Regression Testing is essential for maintaining software quality, but the cost of regression testing is high. Test case prioritization (TCP) is a widely used approach to reduce this cost. Many researchers have proposed regression test case prioritization techniques, and clustering is one of the popular methods for prioritization. The task of selecting appropriate test cases and identifying faulty functions involves ambiguities and uncertainties. To alleviate the issue, in this paper, two fuzzy-based clustering techniques are proposed for TCP using newly derived similarity coefficient and dominancy measure. Proposed techniques adopt grouping technology for clustering and the Weighted Arithmetic Sum Product Assessment (WASPAS) method for ranking. Initially, test cases are clustered using similarity//dominancy measures, which are later prioritized using the WASPAS method under both inter- and intra-perspectives. The proposed algorithms are evaluated using real-time data obtained from Software-artifact Infrastructure Repository (SIR). On evaluation, it is inferred that the proposed algorithms increase the likelihood of selecting more relevant test cases when compared to the recent state-of-the-art techniques. Finally, the strengths of the proposed algorithms are discussed in comparison with state-of-the-art techniques.
Novel Image Compression and Deblocking Approach Using BPN and Deep Neural Network Architecture
(Springer, 2021-04) Viswanathan, Sangeetha
Medical imaging is an important source of digital information to diagnose the illness of a patient. The digital information generated consists of different modalities that occupy more disk space, and the distribution of the data occupies more bandwidth. A digital image compression technique that can reduce an image's size without losing much of its important information is challenging. In this paper, a novel image compression technique based on BPN and Arithmetic coders is proposed. The high non-linearity and unpredictiveness of the interrelationship between the pixels present in the image to be compressed is handled by BPN. An efficient coding technique called Arithmetic coding is used to produce an image with a better compression ratio and lower redundancy. A deep CNN based image deblocker is used as a post-processing step to remove the artefacts present in the reconstructed image to improve the quality of the reconstructed image. The effectiveness of the proposed methodology is validated in terms of PSNR. The proposed method is able to achieve about a 3% improvement in PSNR compared with the existing methods.
Review of Image Fusion Techniques and Evaluation Metrics for Remote Sensing Applications
(Indian Journal of Science and Technology, 2015) Viswanathan, Sangeetha
Low resolution Multispectral images obtained from earth observation satellite interpreted directly have less information whichisnot suitable for remote sensing applications.The advancement of sensors onboardsatellitesprovidespanchromatic imageswhich have more image details and low-resolution MS images lead theway for the researchers to develop algorithms suitable for fusing multi-sensor images which can improve the resolution of the MS images. In this paper review on various pixel based fusion algorithms and evaluation metrics are presented. From the literature review, it is inferred that the Multi resolution techniques will give better accuracy than all other traditional algorithms.
A Scientific Decision Framework for Cloud Vendor Prioritization under Probabilistic Linguistic Term Set Context with Unknown/Partial Weight Information
(MDPI, 2019-05) Viswanathan, Sangeetha
With the tremendous growth of Cloud Vendors, Cloud vendor (CV) prioritization is a complex decision-making problem. Previous studies on CV selection use functional and non-functional attributes, but do not have an apt structure for managing uncertainty in preferences. Motivated by this challenge, in this paper, a scientific framework for prioritization of CVs is proposed, which will help organizations to make decisions on service usage. Probabilistic linguistic term set (PLTS) is adopted as a structure for preference information, which manages uncertainty better by allowing partial information ignorance. Decision makers’ (DMs) relative importance is calculated using the programming model, by properly gaining the advantage of the partial knowledge and attributes, the weights are calculated using the extended statistical variance (SV) method. Further, DMs preferences are aggregated using a hybrid operator, and CVs are prioritized, using extended COPRAS method under the PLTS context. Finally, a case study on CV prioritization is provided for validating the scientific framework and the results are compared with other methods for understanding the strength and weakness of the proposal.
Selection of Apt Renewable Energy Source for Smart Cities using Generalized Orthopair Fuzzy Information
(IEEE, 2020) Viswanathan, Sangeetha
Renewable energy (RE) is a popular and clean source of energy that could potentially reduce carbon footprint and promote sustainable development in smart cities. Developing countries, such as India, have invested time, money, and effort into the proper development of smart cities. As there are different RE alternatives and several criteria used for its selection, researchers have adopted multi-criteria decision-making methods for systematic selection. Previous studies on RE selection did not (i) handle uncertainty effectively; (ii) calculate experts' weights systematically, and (iii) consider interdependencies among experts during aggregation. Motivated by these lacunas, this paper develops a new decision framework. The framework utilizes generalized orthopair fuzzy information, which is flexible and provides rich scope for handling uncertainty. Additionally, a regret theory-based weight calculation method is proposed for systematic weight calculation. Finally, Score-based Muirhead mean is proposed for aggregation of preferences and ranking of REs. An actual case study in Tamil Nadu is presented to exemplify the usefulness of the framework. Comparison with extant models reveals the superiorities of the framework.