Research & Publications

Authors: CIE Research Team

Published in: 2025 8th International Conference on Communication Engineering and Technology (ICCET)

Semantic segmentation is the task of classifying each pixel in the image based on a predefined set of classes. It is a crucial part of an autonomous vehicle to analyze its surroundings and ensure efficient and safe navigation. In these type of applications having low latency and accurate performance is crucial but also challenging since these tasks have high computational complexity. Keeping in mind the resource constraints present in autonomous vehicles, in this study we implement a state-of-the-art segmentation model and evaluate its performance after applying various methods of quantization and pruning to help combat these issues while maintaining high speed and accuracy. We present a comparison of an unoptimized segmentation model along with the different configurations of the optimized model, highlighting the constant accuracy (miou), while providing a 1.61x boost in FPS and a 50% decrease in model size.

Authors: CIE Research Team

Published in: 2025 13th International Symposium on Digital Forensics and Security (ISDFS)

Machine learning serves as an important instrument in financial fraud detection but extreme class imbalance in credit card datasets proves to be a significant obstacle for its effectiveness. Our framework - SURF incorporates two sequential steps where ROA-GAN performs anomaly-based undersampling at its first stage followed by a 1D CNN which carries out fraud classification. Using unsupervised learning SURF efficiently selects critical fraud patterns eliminating the need for synthetic oversampling. This research investigates under-sampling effects on class separation by utilizing PCA and UMAP and benchmarks against Random Under-sampling and NearMiss.

Authors: CIE Research Team

Published in: 2024 IEEE Symposium on Industrial Electronics & Applications (ISIEA)

The existing USB BC1.2 Protocol in contemporary electronic devices relying on USB charging for Li-ion batteries poses a significant challenge today. The dependency on charging for data transfer, dictated by the protocol, results in continuous charging during extended data transfers. This constant charging contributes to accelerated battery wear when there is a requirement for continuous data transfer. This study proposes a charge-controlling module, providing stabilized power input, enabling on-request data access, and actively managing battery health to mitigate degradation. The designed module further employs a battery-capacity-centric algorithm, accommodating diverse device specifications to effectively address economic, safety, and operational concerns associated with continuous charging.

Authors: CIE Research Team

Published in: 2024 IEEE 9th International Conference for Convergence in Technology (I2CT)

The accurate early prediction of the Remaining Useful Life (RUL) of lithium-ion batteries plays an important role for battery health management and predictive maintenance. The insights derived from early RUL prediction can be harnessed to enhance battery lifespan and also for the early detection of potential battery failures, which is particularly critical in the context of the widespread use of lithium-ion batteries in various applications, such as electric vehicles, renewable energy storage, and portable electronics. To address this challenge, we introduce an novel neural network model, the Dual Auto-Encoder Transformer, combined with ensemble techniques for early RUL prediction. Our approach demonstrates a significant improvement in early prediction of the RUL with a Mean Absolute Error (MAE) reduced to 0.0372 and Root Mean Square Error (RMSE) to 0.0451.

Authors: CIE Research Team

Published in: Springer (Chapter in Book "Learning in the Age of Digital and Green Transition")

Remote learning has been in the shadows of mainstream higher education institutions (HEI), with classroom / physical presence taking center stage. The Covid-19 pandemic disrupted the prevalent modes of education and pushed remote learning to the forefront. This project attempts to address some of these critical challenges by building a smartphone application platform for remote learning, emphasizing personalized and gamified learning. The gamified smartphone application with Design Thinking as the first learning module was developed and tested for functionality and usability. This project has received encouraging user feedback and provides a platform for developing and testing more engaging methods in the future for personalized remote learning, using artificial intelligence (AI) and machine learning (ML).

Authors: CIE Research Team

Published in: 2021 IEEE Global Engineering Education Conference (EDUCON)

Research in fast-evolving technologies like AI & ML requires the collaborative effort of various stakeholders including industries and universities. This paper summarizes the IUC effort undertaken by Intel Technology India Ltd and the Center for Innovation and Entrepreneurship at PES University to create mutually rewarding outcomes for both partners and describes a new model encompassing technology, innovation, and entrepreneurship in addition to the traditional elements of IUC. We present the IUC considerations and processes adopted to deal with the challenges and share the outcomes and impact at the end of two years of engagement.

Authors: CIE Research Team

Published in: 2019 IEEE Bombay Section Signature Conference (IBSSC)

This paper presents the development of a cost effective, self-sustainable and solar rechargeable smart farming product based on embedded IoT and providing end to end solution using the cutting edge LoRa technology. It portrays the innovative methods developed to monitor ambient parameters using the LoRa LPWAN. The product leverages the capabilities of the Arduino Mini MCU capitalizing the low power and ultra-low power optimization techniques to enhance self-sustainability of the device. These real time ambient data & parametric trends can be monitored over the cloud at user's convenience on an android application.

Authors: CIE Research Team

Published in: 10th Effectuation Conference (Berlin, 2019)

This paper describes the experience of teaching effectuation as part of an entrepreneurship course for undergraduate technology students offered by the Center for Innovation & Entrepreneurship (CIE) at PES University, India. The experiences are from having taught the course, first offered in Aug of 2018 and followed by Jan 2019 for a combined student strength of 224 across two semesters.

Authors: CIE Research Team

Published in: Academy of International Business: Global Business and Digital Economy, Minneapolis USA

Research presentation at the Academy of International Business conference examining global mindset across BRIC nations, European Union, and emerging breakout economies in the context of digital transformation and global business dynamics.