Autonomous Driving: Object Detection, Path Planning, and Decision Making using Deep Learning Model Rajeshkumar G, MalathiEswaran, K M Subramanian, N Sadhasivam, Vikkram R, S Sadesh Proceedings of 5th International Conference on Soft Computing for Security Applications Icscsa 2025, 2025 In this paper there is an architecture implementable for fault-tolerant pipelines for autonomous driving systems along with path planning, object detection and decision making. The proposed solution is an intermediate-level solution combine several deep learning algorithms in the hierarchical structure, thus improving the reliability and the overall performance of the system. The pipeline begins with generating input data, such as images and videos from dashboard cameras and network streams, and encompasses multiple processing steps up to risk assessment and risk mitigation techniques. For fault tolerance, the architecture relies on fallback techniques since if one algorithm fails, another algorithm prescribed through configuration will be successfully taken over. Road segmentation is achieved using FCNs, but U-Net can also be enabled if necessary. It dynamically deploys YOLO and Single Shot MultiBox Detectors (SSD) for 2D object detection, which allows for algorithm switching corresponding to performance conditions. The object tracking module thus combines Multiple Object Tracking (MOT) algorithms like Deep Sort, Byte Track, and Bot Sort to support accurate and consistent object tracking in dynamic surroundings. To attain thorough situational awareness, the system also uses 3D object detection based on LIDAR. The path planning module ensures safe and effective route navigation by combining well-known algorithms like A*, Dijkstra's Algorithm, and its variations. The multi-layered architecture of the suggested pipeline improves autonomous driving systems' resilience by offering a dependable framework that can preserve operational integrity in a range of circumstances.
Diabetes Diagnosis using Machine Learning Sadhasivam N, Harish J, Bharanidharan M Journal of Trends in Computer Science and Smart Technology, 2023 This abstract presents a study on utilizing the Gradient Boosting algorithm for diabetes diagnosis. The objective is to develop a reliable and effective model that uses patient data, to detect the presence of diabetes. For training and testing, a dataset made up of clinical parameters like age, body mass index, blood pressure, and glucose levels are used. The Gradient Boosting algorithm is implemented and optimized to achieve optimal predictive performance. The model's accuracy, precision, recall, and F1 score are evaluated to assess its effectiveness. The results of this study indicate that the Gradient Boosting algorithm's effectiveness in correctly identifying diabetes and highlight its potential as a trustworthy tool for clinical diagnosis. In order to improve the model's performance and expand its application in real-world healthcare settings, future study can concentrate on adjusting its parameters and investigating new characteristics.
An Analytical Model for Dynamic Spectrum Sensing in Cognitive Radio Networks Using Blockchain Management † Nikhil Kumar Marriwala, Sunita Panda, Chandran Kamalanathan, Narayanan Sadhasivam, Vootla Subba Ramaiah Engineering Proceedings, 2023 Recent advancements in wireless communication technology have brought about the pressing issue of increasing spectrum scarcity. This challenge in spectrum allocation arises from ongoing research in the field of wireless communication. Unfortunately, a significant portion of the spectrum remains underutilized within wireless networks. Cognitive radio (CR) presents an innovative solution to this problem by enabling unlicensed secondary users to coexist with licensed primary users within allocated spectrum bands without causing interference to the primary users’ communications. This paper promises to address the spectrum redundancy challenges and substantially improve the spectrum utilization efficiency. Cognitive radio networks (CRNs), alternatively known as dynamic spectrum access networks, are comprised of multiple CR nodes and are frequently referred to as next generation (XG) communication networks. These XG communication networks are expected to offer high-speed data transmission capabilities to adaptable users through a variety of wireless architectures and dynamic access protocols. Since CRNs share similarities with traditional wireless networks but operate in an external wireless medium, they are more susceptible to various types of attacks compared to their wired counterparts. This vulnerability stems from the fact that wireless media can be intercepted or exploited, potentially leading to channel congestion or data interception. This paper presents two key approaches: the node evaluation and selection (NES) algorithm and the secure spectrum sensing mechanism, which incorporate the user’s interaction history and connection distance, that are recorded in a public ledger and managed by a blockchain management system. The proposed algorithm facilitates the central aggregation point for selecting nodes with outstanding performance for cooperative sensing, thus enhancing the network’s security against malicious node attacks.
Hybrid Genetic Algorithm and Simulated Annealing for Clustering Microarray Gene Expression data M Pandi, T Sivakumar, N Senthil Madasamy, N Sadhasivam Journal of Physics Conference Series, 2021 Gene expression is the process by which information in gene is used to create proteins. The gene expression studies generate large amount of data. These data, referred to as the gene expression matrix, represent the expression levels for thousands of genes recorded at a few time instances. A typical microarray experiment involves the hybridization of an mRNA molecule to the DNA template from which it is originated. Many DNA samples are used to construct an array. The amount of mRNA bound to each site on the array indicates the expression level of the various genes. This number may run in thousands. All the data is collected and a profile is generated for gene expression in the cell. Clustering is a process of partitioning a set of meaningful subclasses called clusters. Clustering is a key step in the analysis of gene expression data. Genetic Algorithms are a family of computational models inspired by evolution. The searching capability of genetic algorithms is exploited in order to search for appropriate cluster center in feature space such that a similarity metric of resulting clusters is optimized. The chromosome which are represented as strings of real numbers, encode the centers of fixed number of clusters. The experiment results are demonstrated on real data sets and the performance of GA is evaluated in comparison with the state-of-the art algorithm K-Means with use of internal validation criteria.
Cancer diagnosis epigenomics scientific workflow scheduling in the cloud computing environment using an improved PSO algorithm N. Sadhasivam, R. Balamurugan, M. Pandi Asian Pacific Journal of Cancer Prevention, 2018 Objective: Epigenetic modifications involving DNA methylation and histone statud are responsible for the stable maintenance of cellular phenotypes. Abnormalities may be causally involved in cancer development and therefore could have diagnostic potential. The field of epigenomics refers to all epigenetic modifications implicated in control of gene expression, with a focus on better understanding of human biology in both normal and pathological states. Epigenomics scientific workflow is essentially a data processing pipeline to automate the execution of various genome sequencing operations or tasks. Cloud platform is a popular computing platform for deploying large scale epigenomics scientific workflow. Its dynamic environment provides various resources to scientific users on a pay-per-use billing model. Scheduling epigenomics scientific workflow tasks is a complicated problem in cloud platform. We here focused on application of an improved particle swam optimization (IPSO) algorithm for this purpose. Methods: The IPSO algorithm was applied to find suitable resources and allocate epigenomics tasks so that the total cost was minimized for detection of epigenetic abnormalities of potential application for cancer diagnosis. Result: The results showed that IPSO based task to resource mapping reduced total cost by 6.83 percent as compared to the traditional PSO algorithm. Conclusion: The results for various cancer diagnosis tasks showed that IPSO based task to resource mapping can achieve better costs when compared to PSO based mapping for epigenomics scientific application workflow.
Cancer detection in microarray data using a modified cat swarm optimization clustering approach M. Pandi, R. Balamurugan, N. Sadhasivam Asian Pacific Journal of Cancer Prevention, 2017 Objective: A better understanding of functional genomics can be obtained by extracting patterns hidden in gene expression data. This could have paramount implications for cancer diagnosis, gene treatments and other domains. Clustering may reveal natural structures and identify interesting patterns in underlying data. The main objective of this research was to derive a heuristic approach to detection of highly co-expressed genes related to cancer from gene expression data with minimum Mean Squared Error (MSE). Methods: A modified CSO algorithm using Harmony Search (MCSO-HS) for clustering cancer gene expression data was applied. Experiment results are analyzed using two cancer gene expression benchmark datasets, namely for leukaemia and for breast cancer. Result: The results indicated MCSO-HS to be better than HS and CSO, 13% and 9% with the leukaemia dataset. For breast cancer dataset improvement was by 22% and 17%, respectively, in terms of MSE. Conclusion: The results showed MCSO-HS to outperform HS and CSO with both benchmark datasets. To validate the clustering results, this work was tested with internal and external cluster validation indices. Also this work points to biological validation of clusters with gene ontology in terms of function, process and component.
Design of an improved PSO algorithm for workflow scheduling in cloud computing environment N. Sadhasivam, P. Thangaraj Intelligent Automation and Soft Computing, 2017 Workflows have been used to represent a variety of applications involving high processing and storage demands. As a solution to supply this necessity, the cloud computing paradigm has emerged as an on demand resource provider. Cloud computing environments facilitate applications by providing virtualized resources that can be provisioned dynamically. User applications may incur large data retrieval and execution costs when they are scheduled taking into account of ‘execution time’ only. In this work, proposed is an Improved Particle Swarm Optimization (IPSO) to schedule applications in cloud resources. The IPSO is used to minimize the total cost of placement of tasks on available resources. Total cost values are obtained by varying the communication cost between the resources, task dependency cost values, and the execution cost of compute resources. Compared with standard PSO, the results show that the improved algorithm is efficient.
