BITS Faculty Publications
Permanent URI for this communityhttp://localhost:4000/handle/123456789/1867
Browse
67 results
Search Results
Item Asynchronous deep reinforcement learning for semantic communication and digital-twin deployment in transportation networks(IEEE, 2025-08) Gupta, ShashankThe dynamically evolving and technologically-driven hybrid landscape of transportation networks integrated with advanced edge computing capabilities has demonstrated efficient communication and computation techniques to guarantee robust quality of services (QoS) to vehicles. However, conventional communication systems in the Internet of Vehicles (IoV) still encounter challenges in providing meaningful low-latency communication and AI-assisted real-time synchronization on the edge. One reason is that it has exhausted the Shannon limit by utilizing cellular, NOMA, and Wi-Fi technologies. Therefore, we present an integrated approach leveraging Semantic Communication (SC), and Digital Twin (DT) deployment to tackle the challenges caused by high-dimensional data exchanges and resource spectrum crunch leading to inevitable latency constraints. SC stimulates meaningful transmission of data to high-mobility vehicles by providing a relevant knowledge base (KB) and DT deployment. In this paper, we established the vehicular SC (VSC) model, and DT deployment strategy. We formulate a multi-objective optimization problem (MOP) to maximize the overall QoS of the system by jointly optimizing VSC and DT deployment. Compared to traditional deep-reinforcement learning (DRL) schemes, we propose a Digital Twin Semantic Sensing using the Multi-vehicle DRL ( DTS2 -MVDL) algorithm which addresses the MOP and persistent issues of multi-dimensional, continuous, and discrete nature of the vehicular environment. Lastly, we employ age of Information (AoI), latency, and QoS as the performance metrics to determine the algorithmic efficiency.Item Leveraging precision agriculture techniques using UAVs and emerging disruptive technologies(Elsevier, 2024-07) Gupta, ShashankThe next great innovation in Unmanned Aerial Vehicles (UAV) technology is smart UAVs, which aim to provide new possibilities in numerous applications. There is an increasing usage of UAVs in various fields of civil applications including live tracking, wireless connectivity, distribution of goods, remote sensing, protection and surveillance, precision agriculture, and review of civil infrastructure. UAVs or drones have a tremen- dous potential to provide smart farming with various productive solutions. Internet of Things (IoT) technologies together with UAVs are anticipated to transform agriculture, allowing decision- making in days rather than weeks, offering substantial cost savings and yield increases. These technologies are employed in a number of different ways, from monitoring crop status and amount of moisture in soil in real time to using drones to help with activities such as the application of pesticide spray. Nonethe- less, the employment of such IoT and smart networking technol- ogy, exposes the smart farming ecosystem to cyber security risks and vulnerabilities. This survey gives a detailed understanding of UAV applications in Precision Agriculture (PA). In this survey, we demonstrate a comprehensive analysis on security and privacy in a smart farming scenario. In this complex and dispersed cyber- physical environment, we describe how Blockchain technology along with 5 G in UAVs communication network can dissipate the security issues of the network. The survey addresses possible scenarios for cyber threats and the advancement in the fields of machine learning and artificial intelligence that can boost cybersecurity. At last, the survey outlines open research issues and future directions in the field of cybersecurity in UAVs and PA.Item Optimizing quality-of-service (QOS) using semantic sensing and digital-twin in pro-dynamic internet of vehicles (IOV)(IEEE, 2024-11) Gupta, ShashankThe emerging autonomous driving industry expects real-time information to be communicated in less amount of time. Most of the extant research works on deterministic or stochastic channels, which are deemed unrealistic for pro-dynamic Internet of Vehicles (IoV) communications. Semantic communication provides a novel concept of serving high-mobility vehicles with faster vehicular communications by using digital twin (DT) technology. However, the low-latency demand, intermittent connectivity, and signal attenuation in the IoV canyon pose big challenges. To facilitate the efficient functioning of Intelligent Transport Systems (ITS) applications, we integrate DT, which is a co-simulation of software such as CARLA, SUMO, python, etc., to improve the semantic communication and quality of service (QoS) of the IoV scenario. Further, we have formulated a vehicular sensing and computation model that incorporates system cost and DT migration cost as their key metrics to evaluate the QoS of the system. We have proposed a pro-dynamic algorithm based on digital-twin deep reinforcement learning (DT-DRL) to decode the QoS maximization problem. Numerical results reveal the superiority of our method by decreasing the cost of the system and improving latency, maintaining the semantic real-time communication in IoV.Item Artificial intelligence inspired task offloading and resource orchestration in intelligent transportation systems(Springer, 2025) Gupta, ShashankInternet of Vehicles (IoV) applications require the support of communication, caching, and computation (3C) resources to offload the computation-intensive tasks and for uplifting the traffic conditions in the development of sustainable smart cities. Intelligent Transportation Systems (ITS) lack the integrated ecosystems of addressing the low-latency task handovers, resource management issues, and centralized incentivization strategies. Digital Twin (DT) aids in capturing the real-time varying resource needs of the vehicles and the communication infrastructure that will regulate the task offloading process and facilitates in incentivizing the vehicular instances. In this manuscript, we establish a digital twin counterpart ( ) of the physical IoV (PIoV) to meet the QoS requirements during dynamic offloading and the time-varying resource supply–demand of computationally intensive applications. We formulate a response delay minimization function which is solved by the proposed DT-driven context-aware dynamic offloading method (CADOM). Furthermore, we use M/M/1/N/FCFS queueing method that combats the drawbacks of handling the simultaneous deadline-based tasks in a volatile environment of PIoV. In addition, we also maximize the utilities of vehicle and RSU service satisfaction by employing a reward-based mechanism for on-demand allocation of resources based on the Stackelberg game, where the DT of vehicle is deemed as a leader and service provider RSUs as a follower. The simulation results establish that the proposed system outpaces the conventional traffic management system by emphasizing the role of in jointly optimizing the overall response latency for different task sizes and also ensure a better utility satisfaction by catering on-demand resource allocation.Item ST-IDS: Spatio-temporal feature-based multi-tier intrusion detection system for artificial intelligence-powered connected autonomous vehicles(Wiley, 2025-03) Gupta, ShashankAdvancements in 3GPP specifications and the extensive deployment of 5G networks have driven significant growth in the Internet of Vehicles (IoVs). This development has led to an increase in Connected and Autonomous Vehicles (CAVs), which provide capabilities such as automated navigation, ADAS, cruise control, and environmentally sustainable transportation in real-time. Additionally, the widespread adoption of CAVs has also escalated vulnerabilities within the IoV ecosystem, exposing it to potential cyberattacks. The integration of various functional interfaces has enlarged its attack surface, thereby increasing the risk of vehicle infiltration. Researchers have proposed various Intrusion Detection Systems (IDS) to address the ongoing risk of vehicle attacks, without applying encryption and related authentication methods for intra-and inter-vehicular communications. However, a significant limitation of many IDSs is their dependency on characteristics specific to a particular category of vehicles, which limits their adaptability. Additionally, current IDSs frequently rely on one-dimensional features such as traffic, time, etc., which limits their capability of detecting attacks in adverse scenarios. Moreover, incorporating machine learning algorithms into IDSs deployed in automated automobiles causes an increase in computational demands. We propose to develop a collaborative IDS specifically designed for cloud-based vehicle environments. We aim to improve our capabilities of identifying intrusion detection and differentiate which are malicious by using multidimensional features. A customised Convolutional Neural Network (CNN), optimised through hyperparameter tuning, is also developed for detecting the malicious vehicles and enhancing the overall IDS. To address the challenge of data diversity, we integrate various vehicular datasets into a unified feature space. This integration allows a single model to efficiently perform multi-classification tasks without frequent adjustments. Our feature space integrates dimensions such as traffic, time and so forth, levels, thereby expanding the spectrum of detectable attack scenarios. By identifying abnormal data points within this comprehensive feature framework, our system effectively identifies intrusions across a diverse range of vehicle types. As a result, our methodology supports robust intrusion detection through comprehensive multiclass vehicle classification.Item HS-GIoV: High-speed green internet of vehicles (IoV) edge-assisted model for low-latency inference in autonomous driving(Elsevier, 2025-08) Gupta, ShashankGreen IoV has emerged as a latent solution in the field of autonomous driving for the future Intelligent transportation system (ITS) accompanied with green wireless communication and computational intelligence. It facilitates enhanced traffic management applications, reduced traffic congestion and compatible V2X connectivity. However, GIoV faces significant challenges in providing seamless bandwidth for real-time video analytics, especially under adverse environments, with improved accuracy in autonomous driving. Although deep neural networks (DNN) are effective in locating vehicles, they struggle to frequently access the edge network and maintain accuracy. In addition, their substantial computational demands waste energy and render them infeasible to deploy on resource-constrained devices for low-latency real-time inference. In this paper, we propose a high-speed GIoV (HS-GIoV) framework that models the problem of energy-efficient video analytics accuracy over multiple time-periods using Lyapunov optimization. To solve this problem, we have proposed a novel on-the-fly Traffic Stream Object Detection (TSOD) algorithm which is lightweight and triggers the re-training only when there is an accuracy decline, thereby avoiding unnecessary computations. We have also proposed a heuristic algorithm that solves seamless bandwidth issue using Lagrangian relaxation. We have tested the HS-GIoV on the self-driving kit that comprises NVIDIA high-end devices. It enhances the accuracy around 20% and reduces the training time to approx. 55%.Item Comparative study of risk assessment models corresponding to risk elements(IEEE, 2012) Gupta, ShashankIn the modern era of software engineering, the development of software in static and dynamic environment results in several vulnerabilities that need to be handled so that they do not step in with the clear defined project goals. Previous studies show that the wide variety of different risk analysis strategies provide a valid solution to address the lack of risk management strategies in Software risk assessment model (SRAM), Software risk assessment and estimation model (SRAEM) etc. In this paper we have discussed the comparison between different software risk assessment models corresponding to certain risk elements. These risk elements must be taken into account in order to cover some perspectives of the software industry which have not been covered up to now. Based on this analysis, we have also concluded the weaknesses and strengths of risk assessment models.Item BDS: Browser Dependent XSS Sanitizer(IGI Global, 2015) Gupta, ShashankCross-Site Scripting (XSS) attack is a vulnerability on the client-side browser that is caused by the improper sanitization of the user input embedded in the Web pages. Researchers in the past had proposed various types of defensive strategies, vulnerability scanners, etc., but still XSS flaws remains in the Web applications due to inadequate understanding and implementation of various defensive tools and strategies. Therefore, in this chapter, the authors propose a security model called Browser Dependent XSS Sanitizer (BDS) on the client-side Web browser for eliminating the effect of XSS vulnerability. Various earlier client-side solutions degrade the performance on the Web browser side. But in this chapter, the authors use a three-step approach to bypass the XSS attack without degrading much of the user's Web browsing experience. While auditing the experiments, this approach is capable of preventing the XSS attacks on various modern Web browsers.Item Efficient Service Utilization in Cloud Computing Exploitation Victimization as Revised Rough Set Optimization Service Parameters(Elsevier, 2015) Gupta, ShashankCloud computing is an effort in delivering resources as a service. In cloud computing setting the role of service supplier is split into two parts as Cloud Broker and repair suppliers. The Cloud Brokermanages cloud platforms and lease resources in keeping with a usage-based evaluation model. The repair suppliers rent resources from one or several infrastructure suppliers to serve the top users. The plan of action of choosing a Cloud Service supplier is evaluated upon the premise of Which-Cloud Provider-Provides-What. Selecting qualification applicableService supplier is more durable as results of all CSPs cannot be counted for all non-stop Service. The aim of this analysis work is to traumatize the programming of the requests on the premise of twelve parameters that got higher best-known to comprehend the simplest best ways that of cloud service supplier allotment to the users. Apart from the implementation and compression purpose taken identical four parameters that unit of measure gift in ROSP recursive program. It uses rough math's to urge the mathematical model inside that the algorithmic program Rough set improvement Service Parameters is created on the premise of the economical resource Utilization in Cloud Computing practice Revised ROSP programming Technique. Then the algorithm is enforced within the cloud machine within that cloudlets, datacenters, and cloud brokers unit of measure wont to perform the algorithms. Some integral packages of Cloud machine unit of measure won’t to simulate the strategy. The strategy is completed combined at a lower place internet Beans and Sql. The results once the implementation of the ERROSP algorithm got unit of measure on high of theROSP algorithm in time taken and mainframe utilization.Item A Combined Model to Ensure Complete Security and Reliability in Cloud Computing(WCECS, 2015) Gupta, ShashankCloud Computing is the fastest growing technique in the IT (Information Technology) industry as its main idea is to maximising the capacity and capabilities vigorously without investing in new infrastructure and licensing software. It provides a large amount of storage capacity over the internet but the management and security of the data and services over the cloud is not entirely trustworthy. Because of the lack in trust, most of the businesses are still reluctant to deploy their business over cloud, so security is the major concern in cloud computing and becoming a major issue in the implementation of cloud. In this paper, a new framework is proposed which focuses on almost every aspect of security ie protection of data from beginning to end, ie, from cloud owner to user. This work focuses on major four aspects of security, ie, Confidentiality, Availability, Integrity and Non-Repudiation. This framework will work on all the categories of Cloud ie Public, Private and Hybrid Cloud and proposes an algorithm to select the correct category of cloud to put a data on to it