MIMO Cognitive Radio Spectrum Sharing Using Spatial Coding and User Scheduling for Fading Channels

Vijayakumar Ponnusamy,Dr. S. Malarvizhi 보안공학연구지원센터 2016 International Journal of Multimedia and Ubiquitous Vol.11 No.3

Spectrum scarcity is one of the issue in wireless communication. Dynamic spectrum sharing is the solution for this problem in which the underutilized spectrums are utilized efficiently. Multiple Input Multiple Output (MIMO) based cognitive radio using spatial coding is used to dynamically share the spectrum which allows simultaneous usage of spectrum by more than one user. The drawback of spatial coding is the achievable capacity get reduced with the increased number of cognitive users which limits the number of users to share the spectrum. In order to maximize the number of secondary users sharing the spectrum of the primary user, this paper presents a cluster based multiple primary and secondary user’s MIMO spatial coding which provides higher capacity. The entire network is divided into number of cluster based on the location such that each cluster will be interference free from other. Each cluster is designed with one primary user and many secondary users and a User Scheduling algorithm called Cluster Based Max Signal Power User Scheduling (CBMSPUS) is proposed to select two best secondary users in each cluster. Since the performance of spatial coder depends on the channel matrix, this paper analyze the spatial pre-coder under Rayleigh, Rican and Nakagami fading channels with multiple users and evaluate the capacity of users under CBMSPUS scheduling algorithm. The results show that the Nakagami fading channel achieves high capacity than others. Primary user calculated capacity is 10.8bits/HZ /channel at low SNR region and 14.2 bits/HZ/channel at high SNR region. The secondary user archives maximum capacity of 9.8bits/HZ /channel at low SNR and 13 bits/HZ/channel at high SNR. The Proposed Scheduling method offers maximum of 15dB gain comparing FIFO scheduling algorithm.

A Palm Vein Recognition System based on a Support Vector Machine

Vijayakumar Ponnusamy,Abhijit Sridhar,Arun Baalaaji,M. Sangeetha 대한전자공학회 2019 IEIE Transactions on Smart Processing & Computing Vol.8 No.1

Palm vein authentication is among the recent research in the field of access control applications. Because palm veins are tough to forge, they act as a reliable metric in security applications. Accurate extraction of palm veins is challenging in the presence of various dynamics, such as variant light conditions, variations in palm vein patterns from person to person, the cleanliness of the hand, etc. This paper presents a robust recognition process that makes use of a ridge filter for vein pattern extraction, and local binary patterns (LBP) for feature extraction. The ridge filter takes the major eigenvalue of the Hessian matrix, which contains the second-order derivative of the image pixels. The eigenvalue is then processed using LBP feature extraction from the vein patterns. Finally, a support vector machine is used for classification of subsequent images. The result shows that the system can provide accuracy of 89%, with a computation time of 0.423s. The false acceptance rate and false rejection rate were also evaluated as benchmark parameters, which show significantly good performance.

Tuning of a PID Controller Using Soft Computing Methodologies Applied to Basis Weight Control in Paper Machine

Balakrishnan Nagaraj,Ponnusamy Vijayakumar 한국펄프·종이공학회 2011 펄프.종이기술 Vol.43 No.3

Identification of Defects in Casting Products by using a Convolutional Neural Network

Dilliraj Ekambaram,Vijayakumar Ponnusamy 대한전자공학회 2022 IEIE Transactions on Smart Processing & Computing Vol.11 No.3

The main perspective when ensuring dependability in speculations over accuracy in casting parts is a project quality confirmation process that is both careful and meticulous under Industry 4.0. When thorough and extensive casting project examination strategies merge with expanded metal project quality standards, casting production, augmented visual inspections, ensemble process modification and execution are improved. In this paper, we use publicly available casting image datasets for visual inspection, which classify defective and non-defective casting. Inspired by the convolutional neural network (CNN), we propose two-stage convolution for modeling, with DenseNet for classifying casting products. Through experimentation, we achieved an F1-score of 99.54% with a processing time of 454ms using a CPU for classification of casting product inspections. The modified modeling of the CNN in this work helps to improve optimization, compared to other basic machine learning mechanisms that measure quality.

AI-assisted Physical Therapy for Post-injury Rehabilitation: Current State of the Art

Dilliraj Ekambaram,Vijayakumar Ponnusamy 대한전자공학회 2023 IEIE Transactions on Smart Processing & Computing Vol.12 No.3

Telemedicine in physical therapy has increased rapidly since the COVID-19 pandemic erupted. Recuperation can help survivors resume their lives by restoring lost skills, regaining independence, and improving their well-being. With the help of innovative technologies, researchers have created new methods to aid clinicians in patient evaluation and assessment, and more people than ever have access to physiotherapy. The focus of this study is the use of deep learning and machine learning algorithms in conjunction with virtual, augmented, and mixed reality (VR, AR, and MR) technologies for experimental analysis to help patients recover from intracranial hemorrhage, stroke, musculoskeletal and neurological trauma, scoliosis, etc. We present evaluation frameworks systematically categorized into three groups: detecting emotions, identifying movements, and mimicking clinical assessments. We also examine the most popular sensors, body regions, and outcome metrics, and we review plans in evaluating AI strategies (from element design to grouping). Finally, some challenges and future directions for reviewing the field are presented.

An IoT based Cloud EEG Signal Analytic Framework for Thought to Text Mapping

A. Joshua Jafferson,Vijayakumar Ponnusamy,Jovana Jovic,Miroslav Trajanovic 대한전자공학회 2021 IEIE Transactions on Smart Processing & Computing Vol.10 No.3

Paralyzed people have difficulty communicating with the world for their daily basic needs, and their caretakers have difficulty understanding their needs. The development and implementation of a handheld device-based brain-computer interface system with machine learning will solve the above problem. On the other hand, a simple handheld device cannot satisfy the computation of hunger ML algorithms and will have more latency. This paper overcomes the limitations of the above by processing the data in the cloud. The handheld device reads and preprocesses the electroencephalogram (EEG) data and forwards it to the IoT-based Cloud server. The cloud server applies the machine-learning algorithm and classifies it in the text, representing the word thought by the user. This text information result is sent back to the handheld device and intimates the caretaker to know the patient"s needs. The evaluation result of the proposed system for ten words to deal with the basic needs highlights the feasibility of implementing it in practice.

Diagnosis of Non-invasive Glucose Monitoring by Integrating IoT and Machine Learning

V. K. R. Rajeswari Satuluri,Vijayakumar Ponnusamy 대한전자공학회 2022 IEIE Transactions on Smart Processing & Computing Vol.11 No.6

Diabetes Mellitus (DM) is a term collectively used for all types of diabetes. DM increases the risk factor for health complications if not treated early. The Internet of Things (IoT) and artificial intelligence (AI) in healthcare have become a huge benefit for managing DM. The selfsupervision of healthcare has become convenient because of IoT-enabled devices. This paper reviews the management of diabetes, such as invasive, non-invasive, and minimally invasive methods. Justification for the need for non-invasive monitoring of glucose is discussed. Different AI and IoT-enabled management for non-invasive diabetes are also briefed. This review aims at the type of machine learning algorithms applied to non-invasive glucose monitoring. The following are to be considered to achieve an effective non-invasive method of monitoring glucose: Near Infrared spectroscopy (NIR) and Machine learning algorithms(ML). IoT in glucose monitoring has empowered doctors and caretakers to deliver outstanding care. Self-care by every person has become essential, which can be achieved by handheld or wearable IoT devices. Using current technologies, the possibility of making a wearable to monitor the glucose level is becoming closer to reality and has enormous potential.