Journal of IoT Security and Smart Technologies (e-ISSN: 2583-6226)

How Machine Learning Helps in Privacy and Protection

Ritu Kaushik — 2026-05-30

In today’s world, privacy and protection-related issues are becoming increasingly relevant day by day. It becomes a critical concern due to the rapid growth in the usage of social media platforms, internet usage, and smart usage of technology, along with its advancements. Individuals continuously share their personal data or information through online services. Individuals connect with one another with online networking platforms, due to which there is always a threat of data breaches, identity theft, unauthorized surveillance and phishing, etc. With the increasing use of Artificial Intelligence and Big Data Analytics, these issues are becoming more intense day by day, enabling the collection of large-scale datasets. This paper examines the causes of the increase of major privacy concerns like lack of awareness, excess data tracking and weak security mechanisms in digital systems. The study explores how machine learning techniques can be used to enhance data protection and ensure secure data processing. The paper concludes that achieving a balance between technological innovation and privacy protection is essential for building a secure and trustworthy digital environment in the future.

A Novel AI IoT-based Approach for Autonomous Waste Segregation and Management in Sustainable Smart Urban Environments

K. Murugan — 2026-06-06

Rapid urbanization has resulted in a considerable rise in municipal solid waste, creating significant environmental and public health issues in contemporary cities. Conventional waste management systems frequently prove to be inefficient, labor-intensive, and deficient in real-time monitoring capabilities, leading to improper waste segregation and resource wastage. To overcome these challenges, this paper introduces an innovative AI-IoT (Artificial Intelligence of Things) strategy for autonomous waste segregation and management in sustainable smart urban settings. The suggested system combines Internet of Things (IoT) sensors with Artificial Intelligence (AI) algorithms to facilitate real-time detection, classification, and sorting of waste into categories such as biodegradable, recyclable, and hazardous materials. Sophisticated machine learning models, including computer vision techniques, are utilized to accurately identify different types of waste, while embedded sensors keep track of bin levels, environmental conditions, and collection schedules. Furthermore, the system integrates cloud-based data analytics to optimize waste collection routes, lower operational costs, and improve decision-making for urban authorities. Additionally, the framework fosters sustainability by enhancing recycling efficiency, decreasing reliance on landfills, and minimizing environmental pollution. The autonomous characteristics of the system lessen the need for human intervention, ensuring safer and more hygienic waste handling processes. Experimental analyses and simulations indicate that the proposed AIoT-based solution markedly enhances segregation accuracy, collection efficiency, and overall waste management performance when compared to traditional methods. This research aids in the advancement of intelligent, scalable, and eco-friendly waste management systems, aligning with the vision of sustainable smart cities.

Project Desk: An Automated Role-based Project Allocation and Management System using MERN Stack

Sumit Kumar — 2026-08-10

Project allocation and monitoring are integral components of project-based learning and play an important role in the college experience. However, managing these processes can be time-consuming and administratively demanding. In many institutions, activities such as team formation, mentor allocation, and progress tracking continue to rely on spreadsheets and manual documentation, resulting in inefficient and fragmented processes. Managing hundreds of students and dozens of deadlines can make it easy for important tasks to be overlooked, leaving both students and faculty feeling disconnected and overwhelmed. This paper presents Project Desk, an automated, role-based web application developed using the MERN stack (MongoDB, Express.js, React, and Node.js) to streamline the complete project allocation lifecycle. The system provides separate dashboards for students, project coordinators, and mentors, enabling structured workflows, secure authentication using JSON Web Tokens (JWT), and efficient document handling through Multer. Students can propose new project ideas or select from a project bank, coordinators can review proposals and schedule interviews, and mentors can monitor progress and provide feedback. The proposed system improves efficiency, reduces manual effort, enhances transparency, and ensures systematic project monitoring. Experimental evaluation shows that Project Desk significantly outperforms traditional manual methods in terms of time management, communication efficiency, and data organization.

A Lightweight Authentication Protocol for Wearable IoT Healthcare Devices: Design, Security Analysis, and Performance Evaluation

Shikha Tiwari — 2026-08-13

Wearable Internet of Things (IoT) devices used for continuous health monitoring — such as ECG patches, glucose monitors, smart insulin pumps, and fitness bands — transmit highly sensitive physiological data over resource-constrained wireless channels. Conventional public-key-based authentication schemes (RSA, bilinear pairing) impose computation, memory, and energy costs that exceed the capability of these battery-limited devices, exposing patients to impersonation, replay, and man-in-the-middle attacks. This article proposes a lightweight, hash- and XOR-based mutual authentication protocol tailored to wearable IoT healthcare devices, in which a Gateway Node (GWN) mediates registration, mutual authentication, and dynamic session-key agreement between the wearable sensor and the remote healthcare server, without relying on expensive asymmetric cryptographic primitives. The protocol further embeds a Physically Unclonable Function (PUF) response at registration for device-level anti-cloning protection and enforces periodic pseudo-identity and session-key refresh to defeat traceability attacks. Formal and informal security analysis shows resistance to replay, impersonation, man-in-the-middle, desynchronization, and stolen-verifier attacks while preserving anonymity and forward secrecy. Simulation-based performance evaluation demonstrates that the proposed protocol reduces computation time by up to 78% and energy consumption by up to 73% relative to RSA and ECC-based baselines, while incurring lower communication and storage overhead, making it well suited for large-scale, battery-powered wearable healthcare deployments.

Agentic Honeypot Intelligent API Interaction Monitoring System

Tulasi K. L. — 2026-08-17

Cybersecurity threats have become increasingly sophisticated due to the rapid growth of web applications, cloud computing, Application Programming Interfaces (APIs), and automated attack tools. Recent advances in Artificial Intelligence (AI), Machine Learning (ML), Large Language Models (LLMs), and intelligent cyber-deception technologies have enabled the development of adaptive honeypot systems capable of interacting dynamically with attackers. These systems employ attack monitoring, malicious payload analysis, shell emulation, attacker behavior modeling, LLM-generated responses, and adaptive deception strategies to collect detailed threat intelligence while minimizing exposure of real computing resources. This survey reviews recent research on intelligent honeypot systems, focusing on traditional interaction-based honeypots, machine-learning-based approaches, LLM-powered honeypots, adaptive attack modeling, and intelligent API interaction. The study compares existing approaches, identifies their strengths and limitations, and discusses future research directions toward developing secure, adaptive, and intelligent agentic honeypot systems for monitoring malicious API interactions.