http://localhost:8081/jspui/handle/123456789/116 2026-10-21T22:23:02Z 2026-10-21T22:23:02Z Aseeja, Ashish http://localhost:8081/jspui/handle/123456789/21627 2026-09-21T10:25:06Z 2023-05-01T00:00:00Z

Title: Social Media Analysis to Identify Friend or Foe Authors: Aseeja, Ashish Abstract: The Data generated online is increasing at a breakneck pace. With over 315 million monthly active users and over 502 million tweets per day, Social Media platforms have become a source of significant data generation that, if analyzed, can be used to prevent cyber-attacks. Social media platforms are used by enemy agencies and banned organi zations for attacking defense personnel. Identifying these personnels on social media will help safeguard assets and avoid information lapses. Using manual searches, this identification of profiles is laborious, manpower-consuming, and inefficient. Identifying friend or foe on social media can be complex due to the inherent nature of online interactions, where individuals may hide their true identities and motivations. However, some certain strategies and indicators can help users distinguish between gen uine friends and potential adversaries. One area where social media, particularly Twit ter, has made a significant impact is sentiment analysis. Sentiment analysis involves the process of gauging the emotions, opinions, and attitudes expressed in text, in this case, tweets. The aim of this research is to automate the process of identifying friend or foe from the tweets extracted using Twitter.Identifying Friend or Foe is based on identifying excessive positive or excessive negativity and aggression in online interactions. In this research, it is shown that the users can be clustered into friend or foe based on the contents of the tweets and the same can be used for any trending topic in general. These users can then be monitored for their activities on social media, which can help prevent social media crimes.

2023-05-01T00:00:00Z Bhardwaj, Chaitanya http://localhost:8081/jspui/handle/123456789/21626 2026-09-21T10:24:52Z 2023-06-01T00:00:00Z

Title: Beamforming for Cell-Free Downlink Massive MIMO Systems Authors: Bhardwaj, Chaitanya Abstract: With the release of fifth-generation (5G) wireless communication networks, we can experience a significant improvement in reliability, data rate, and energy efficiency. These improvements are achieved by a key technology called massive multiple-input multiple-output (mMIMO). The mMIMO system exploits the benefits of MIMO sys tems by deploying a large number of antennas at the base station (BS) and concurrently serving a large number of users. As we advance, there is a need for ubiquitous and seamless connectivity to support several existing and new applications. One candi date technology that can facilitate such connectivity is cell-free massive MIMO (CF mMIMO). In CF-mMIMO, hundreds of access points (APs) are deployed in a large geographical, where a set of APs beamform the signal to the intended user and enhance the received signal-to-interference-noise ratio (SINR). Since a user is supported by a set of APs, the beam weights at the AP antennas have to be optimized suitably. In this work, westudy the existing beamforming optimization techniques for CF-mMIMO sys tems, from the basic ones like equal power allocation (EPA) to the advanced ones like weighted minimum mean squared error (WMMSE). We observe that WMMSE has su perior performance but requires a higher computational complexity. Thus, we propose using the interior point method (IPM) to solve the beamforming optimization problem. Using simulations, we corroborate that the proposed IPM-based beamforming requires lesser time complexity and converges faster than the existing WMMSE algorithm at the cost of a marginal drop in the achievable throughput.

2023-06-01T00:00:00Z Singhal, Manmohan http://localhost:8081/jspui/handle/123456789/21625 2026-09-21T10:24:39Z 2023-06-01T00:00:00Z

Title: Joint Detection and Cancellation of Cross-Interference in Automotive FMCW Radars Authors: Singhal, Manmohan Abstract: Frequency Modulated Continuous Wave (FMCW)radars have emerged as a promis ing sensing modality for implementing Advanced Driver Assistance Systems (ADAS) for autonomous vehicles due to their capability of all-weather, precise, and high-resolution detection of objects. Dense deployment of FMCW radar equipped autonomous vehicles in urban scenarios can increase the risk of cross-interference between multiple radars operating in close proximity. Cross-interference leaking into the received baseband sig nal of the victim radar can increase its perceived noise floor which can significantly degrade the detection of weak objects/obstacles. In this thesis, we study the structure of cross-radar interference and its impact on detection of weak objects. We propose an algorithmic framework for canceling the cross-interference in the received baseband signal at the victim radar. We develop a novel structure-aware proximal subgradient descent algorithm to blindly separate the interference-free received baseband signal consisting of complex sinusoids and the temporally-sparse cross-interference signal. We conduct experiments using both simulated and real-world FMCW radar measure ments to demonstrate the accurate, robust and real-time cross-interference cancellation performance of our proposed algorithm compared to existing interference mitigation methods. Lastly, we also discuss the techniques to accelerate the convergence and fur ther reduce the runtime of our proposed interference cancellation algorithm.

2023-06-01T00:00:00Z Bo, Myo Chit http://localhost:8081/jspui/handle/123456789/21624 2026-09-21T10:22:47Z 2023-06-01T00:00:00Z

Title: FORENSICS OF JPEG IMAGES Authors: Bo, Myo Chit Abstract: Social networking platforms extensively employ digital images for sharing visual information. The accessibility of user-friendly image editing software has facilitated the creation of convincingly authentic counterfeits. In order to validate the contents of digital images even in the face of adversary activities, it is therefore vital to create strong forensic tools. JPEG formats have achieved substantial popularity, establish ing themselves as one of the most extensively utilized image formats. In the field of digital image forensics, having a thorough understanding of the JPEG format’s funda mental properties can be of great value. The three main artifacts associated with JPEG compression are quantization, rounding, and truncation. We investigate how these ar tifacts affect JPEG forensics through theoretical research. We investigate anti-forensic methods that effectively gets rid of the blocking artifacts of earlier JPEG compression in an image. Furthermore, we demonstrate the practical usefulness of this method in deceptive forensic methodologies explicitly developed to determine image origins, de tect instances of double JPEG compression, and find instances of manipulated image forgeries using cut-and-paste methods.

2023-06-01T00:00:00Z