About

Ranjitha Kumar is an Associate Professor at the Siebel School of Computing and Data Science at the University of Illinois Urbana-Champaign. Her research areas include Interactive Computing, User-Centered Machine Learning, and Human-Computer Interaction. She has taught courses such as The Art of Web Programming, User-Centered Machine Learning, and HCI for ML, and is involved in initiatives related to data-driven design and innovation in computing. Ranjitha Kumar serves as the Director of the Innovation Leadership and Engineering Entrepreneurship (ILEE) Program in the Grainger College of Engineering, contributing to the development of leadership and entrepreneurial skills within engineering education.

Research topics

• Computer Science
• Artificial Intelligence
• Machine Learning
• Computer engineering
• Embedded system
• Electrical engineering
• Parallel computing
• Computer hardware
• Computer architecture
• Engineering
• Programming language
• Theoretical computer science

Selected publications

• Multi-layer Security and Power Efficiency Improvement with Blockchain Technology and NB-IoT

  Research Square · 2024-04-10

  preprintOpen access

  <title>Abstract</title> To increase the level of safety, precision, and openness of narrow-band IoT combined with blockchain. NB- IoT revolutionizes seamless connectivity for IoT devices in 3Gpp released 15, enabling long battery life and widespread coverage, its low-power consumption, enhanced penetration capabilities, promising a scalable and robust infrastructure for the future with integration of Blockchain technology presents a decentralized, immutable ledger system that ensures transparent and secure transactions across various industries, fostering trust without intermediaries and reducing fraud risks significantly. Its cryptographic principles and distributed nature empower industries by enabling traceability, enhancing data integrity, and revolutionizing traditional processes of supply chain management, our approach provides distributed data authentication while simultaneously creating a layer-based blockchain with NB- IoT proposing in this paper 16 QAM OFDM modulation technique optimizing Bit Error rate (BER), Symbol Error rate (SER), and Frame Error rate (FER) signal to noise ratio (SNR) at the maximum power of 25db using MAT Lab simulation graphical results in multi-layer, double-layer, and single-layer. Our mathematical model for optimizing business activities carried out at every stage of the intelligent supply chain facilitates revenue sharing among the entities of producers, distributors, and re-sellers who make up the proposed blockchain system and NB-IoT as it incorporates with additional traceability and review functions, it is the optimal solution for safe transportation of intelligent supply chain management.

  PublisherOA PDFDOI

• 5G'S Consequences on Industry Digital Evolution and Automation

  2024-09-20

  article1st authorCorresponding

  The fifth-generation connection is being developed and implemented by the mobile industry. One of the main factors propelling the development of IoT and other intelligently managed applications is the growing availability of 5GHz relationships. Blockchain, the World of Everythings, artificially intelligent driverless vehicles, accessible reality, and a host of other breakthroughs that we haven't yet begun to imagine are all dependent on 5GHz's blazing-fast connectivity and low latency. Beyond just marking a shift in generations, 5GHz offers new opportunities for all areas that are reachable. This study aims to evaluate relevant material and investigate how 5GHz could facilitate or expedite smart robotization across a range of sectors. This essay looks at how future generations of mobile annex devices will grow, emphasizes the significance of the 5GHz relationship, investigates the technologies that support them, looks at their trends and obstacles, investigates their applications in various manufacturing sectors, and emphasizes the influence they will have on the era of endless connectivity, intelligent tools, and accessible manufacturing.

  PublisherDOI

• A Way of Optimization of Wireless Sensor Network using TSCH

  2024-05-14 · 24 citations

  article

  Within ad hoc wireless networks, TimeSynchronized Channel Hop (TSCH) is a potential method for interference-free and collision-free communication, especially in the setting of wireless sensor networks, car networks, and networks containing robots or drones. Even though TSCH and its versions work well, they need centralized planning to keep the time-frequency sliding system in place. This leads to a restricted usage of time-frequency slots and slow convergence, especially when there is node change or movement. Using the idea of pulse-coupled oscillators, the protocol pairs distributed synchronization or desynchronization methods at the MAC layer to coordinate nodes to perform coordinated signal packet broadcasts over all available channels. Afterwards, peer-to-peer swap requests and acknowledgements between ongoing emitter in neighboring channels result in autonomous channel moving. Extensive evaluation against TSCH and the power source Successful It MAC (EM-MAC) protocol suggests that DT-SCS is a strong contender for centralized multi-channel MAC layer coordination due to its quick convergence, which refers strong connectivity, and high bandwidth utilization even in the midst of interference alongside hidden nodes.

  PublisherDOI

• Reliability and Validity of Questionnaire for Adhesive Capsulitis

  African Journal of Biomedical Research · 2024-12-30

  articleOpen access1st authorCorresponding

  Adhesive capsulitis (AC), also known as frozen shoulder, is a condition characterized by pain and significant loss of both active range of motion and passive range of motion of the shoulder. It usually affects patients aged 40-70 years, with females affected more than males, and no predilection for race. The prevalence of adhesive capsulitis is estimated to be 2% to 5% of the general population and higher incidence of frozen shoulder among patients with diabetes (10-20%), there is an even greater incidence among patients with insulin dependent diabetes (36%), with increased frequency of bilateral shoulder involvement. The available questionnaires take in to account the constructs of pain and/or functional limitations of the shoulder but do not assess psychological or social problems related to the conditions of the shoulder. Till date no questionnaire specific to AC has been found in literature, therefore, the objective of this study is to design a comprehensive questionnaire to assess prevalence, risk factors, impact and healthcare utilization among subjects with adhesive capsulitis. The content validity and reliability of the questionnaire was established. Items showing acceptable internal consistency, test-retest reliability are considered in the final questionnaire. The questionnaire is ready with necessary amendments and will be used on larger sample size in main study.

  PublisherOA PDFDOI

• A Novel Hybrid Clustering Method for Secured Cross-Layer and Cross-Domain Routing in Dense Wsns

  2024-12-13

  article

  Discovering a method that would allow the networks to last for a longer period of time is one of the utmost grave problems in Wireless Sensor Networks (WSN). It is necessary to model method that is efficient in order to conserve limited power assets that are available for the Wireless Sensors Network. One solution to these issues is the use of cross-layer rules, which are designed to improve the efficiency of routing while additionally extending the lifetime of the network. An innovative model that allows the routing of data across layers and domains is revealed, and subsequently the model is optimized for use in safe clustered inside vast wireless networks of sensor. A unique hybrid technique that combines features of Moth Flame Optimization (MFO) and Dragonfly Algorithm (DA) is presented in this study. The purpose of this strategy is to improve the effectiveness of a density wireless sensor. A comparison of the effectiveness of the suggested model with five other standard techniques is presented in the last section. We examine the impacts on the network over a lengthy period of time, as well as the analysis of living nodes, divided into three distinct groups: normal nodes, evolved nodes, and superior nodes. In terms of both the assessment of whether or not a node is alive and the longevity of the network, we discover that the technique that was provided gives satisfactory results. When compared to Adaptive Lion Optimization Protocol (ALOP), Genuine Grey Wolf Search Optimization (GGWSO), and Adaptive Whale Optimization Algorithm (AW-WOA), the Moth Flame Integrated Dragonfly Algorithm (MFI-DA) technique proposes achieves the highest level of accuracy, which is 96%.

  PublisherDOI

• An Extensive Examination of Different Phases Available in Enhancing the Power of WSN

  2024-05-14

  article

  In this whole test, we delve into the evolution and transformative ability of Wireless Sensor Networks (WSNs), emphasizing the pivotal position of graphical statistics in elucidating the technological improvements and stressful conditions interior this region. By harnessing the energy of the IEEE 802.15.4 general, this research provides an in-intensity analysis of WSNs’ operational efficiencies, energy consumption styles, and network resilience through meticulously designed graphs and information visualizations. These graphical representations now not only offer a easy and concise overview of WSN universal performance metrics but additionally spotlight the giant enhancements completed thru the implementation of a singular WSN framework. The examiner’s findings underscore the vital importance of empirical statistics visualization in advancing our knowledge of WSN technologies, offering valuable insights into the optimization strategies that may result in extra power-inexperienced, reliable, and scalable sensor networks. This research contributes to the continued communicate in the subject, paving the manner for future enhancements and programs of WSNs across several sectors.

  PublisherDOI

• Exploiting Short Application Lifetimes for Low Cost Hardware Encryption in Flexible Electronics

  2023-04-01 · 4 citations

  articleSenior author

  Many emerging flexible electronics [1] applications require hardware-based encryption, but it is unclear if practical hardware-based encryption is possible for flexible applications due to stringent power requirements of these applications and high area and power overheads of flexible technologies relative to silicon CMOS technologies. In this work, we observe that the lifetime of many flexible applications is so small that often one key suffices for the entire lifetime. This means that, instead of generating keys and round keys in hardware, we can generate the round keys offline, and instead store these round keys directly on the engine post fabrication in an on-chip programmable read-only memory. This eliminates the need for hardware for dynamic generation of round keys, which significantly reduces encryption overhead, while still allowing engines to have unique keys. This significant reduction in encryption overhead allows us to demonstrate the first practical flexible encryption engines. To prevent an adversary from reading out the stored round keys, we scramble the round keys before storing them in the ROM; camouflage cells are used to unscramble the keys before feeding them to logic. In spite of the unscrambling overhead, our encryption engines consume 27.4% lower power than the already heavily area and power-optimized baselines, while being 21.9% smaller on average.

  PublisherDOI

• Machine learning guided thermal management of Open Computing Language applications on CPU‐GPU based embedded platforms

  IET Computers & Digital Techniques · 2022-12-28 · 3 citations

  articleOpen access1st authorCorresponding

  Abstract As embedded devices start supporting heterogeneous processing cores (Central Processing Unit [CPU]–Graphical Processing Unit [GPU] based cores), performance aware task allocation becomes a major issue. Use of Open Computing Language (OpenCL) applications on both CPU and GPU cores improves performance and resolves the problem. However, it has an adverse effect on the overall power consumption and the operating temperature of the system. Operating both kind of cores within a small form factor at high frequency causes rise in power consumption which in turn leads to increase in processor temperature. The elevated temperature brings about major thermal issues. In this paper, we present our investigation on the role of CPU during execution of GPU specific application and argue against running it at the high frequency. In addition, a machine learning guided mechanism to predict the optimal operating frequency of CPU cores during execution of OpenCL GPU kernels is presented in this study. Our experiments with OpenCL applications on the state of the art ODROID XU4 embedded platform show that the CPU cores of the experimental board if operated at a frequency proposed by our Machine Learning‐based predictive method brings about 12.5°C reduction in processor temperature at 1.06% degradation in performance compared to the baseline frequency (default performance frequency governor of the embedded platform).

  PublisherOA PDFDOI

• Application driven routing for mesh based Network-on-Chip architectures

  Integration · 2022-01-08 · 14 citations

  article
  PublisherDOI

• Efficient Test Case Generation in Software Testing Using DistilGPT-2 and EfficientNet-Lite

  International Journal of Multidisciplinary Research and Explorer · 2022-02-28

  articleOpen accessSenior author

  Software testing plays a critical role in ensuring software reliability, yet traditional test case generation approaches often suffer from high computational overhead and inefficiency. Traditional methods, including genetic algorithms, struggle with scalability and fail to optimize execution time while maintaining high test coverage. To address these limitations, this paper proposes a lightweight deep learning-based test case generation approach using DistilGPT-2 and EfficientNet-Lite. Unlike conventional deep learning models, our method efficiently generates both text-based and GUI-based test cases while reducing computational cost. The novelty of this approach lies in integrating CodeT5-Small for feature extraction, DistilGPT-2 for textual test case generation, and EfficientNet-Lite with an RNN for GUI-based testing, enabling a more effective, low-resource test generation pipeline. The results demonstrate that our method achieves higher test coverage (95%), improved efficiency (90%), and greater testing reliability (98%) compared to advanced genetic algorithms, while also reducing computational overhead to 60%. Compared to existing approaches, our method outperforms traditional AI-based testing solutions in terms of accuracy, fault detection rate, and efficiency. The proposed method enhances software testing by minimizing redundant test cases, improving execution pass rates, and ensuring broader code coverage, making it a scalable and cost-effective solution for modern software development. This work paves the way for lightweight transformer-based models in test case generation, ensuring robust test automation with minimal resource consumption.

  PublisherOA PDFDOI

Recent grants

• Collaborative Research: Variability-Aware Software for Efficient Computing with Nanoscale Devices

  NSF · $600k · 2010–2016

• Collaborative Research: Software Canaries

  NSF · $90k · 2013–2017

• SHF: Small: Printed Computer Systems

  NSF · $600k · 2020–2024

Frequent coauthors

• John Sartori

  University of Minnesota System

  41 shared

• Dean M. Tullsen

  23 shared

• Henry Duwe

  Iowa State University

  19 shared

• Hari Cherupalli

  Synopsys (United States)

  12 shared

• Weidong Ye

  Fudan University Shanghai Cancer Center

  10 shared

• Norman P. Jouppi

  Google (United States)

  10 shared

• Xun Jian

  Virginia Tech

  8 shared

• Nathaniel Bleier

  University of Illinois Urbana-Champaign

  7 shared

Labs

• Siebel School of Computing and Data SciencePI

Awards & honors

• Celebration of Excellence 2021
• Celebration of Excellence 2022
• Celebration of Excellence 2023
• Celebration of Excellence 2024
• Celebration of Excellence 2025