Research Projects

The proposed research project aims to address the critical need for continuous, real-time, and patient-specific monitoring of epileptic seizures. Epilepsy affects over 50 million people worldwide, and nearly 30% of these individuals do not respond to current pharmacological treatments. Consequently, there is an urgent requirement for automated, reliable, fast and power-efficient seizure detection systems that can operate in real-time, particularly in wearable or implantable medical devices. The scientific objective of this project is to develop and prototype seizure detection system leveraging an improved, hardware-optimized Empirical Mode Decomposition (EMD) technique integrated with a support vector machine (SVM) classifier on an FPGA platform. As compared to the traditional software-based approaches or hardware designs that employ discrete wavelet transform (DWT) or fast Fourier transform (FFT), the proposed design utilizes a data-driven, adaptive EMD technique that is inherently more suited to capture the non-linear and non-stationary nature of EEG signals. This technique, combined with FPGA-specific optimizations, will allow real-time processing with low latency, minimal power consumption, and high classification accuracy. The proposed EMD design, will be implemented using optimized digital logic components such as radix-8 CORDIC modules, extrema detection module, arithmetic modules, FSMs, streaming datapaths etc. when combined with a fixed-point, low-latency SVM classifier, can achieve better seizure detection performance (in terms of sensitivity, specificity, and false alarm rate) as compared to existing methods. To realize this objective, the project carries out several core experiments. These include the development of Verilog-HDL modules for the key functional blocks such as EMD, statistical feature extraction, and SVM classification. The design undergoes simulation and synthesis using Xilinx Vivado tools on FPGA platforms like Artix-7/Kintex-7/Zynq-7000. Performance benchmarking is conducted using both benchmark datasets such as the CHB-MIT Scalp EEG Database, and real-time EEG signals collected from Brahmananda Narayana Multispeciality Hospital, Jamshedpur. The system is evaluated based on critical design metrics including detection latency, power consumption, and classification accuracy under realistic clinical and hardware constraints. After completion of the project, the project will result in a deployable seizure detection module that is not only scientifically robust but also practically scalable for the real-world use. It holds high significance in the domain of biomedical engineering, as it bridges the gap between algorithmic innovations in signal processing and real-time embedded medical applications. The developed system will enhance clinical safety, support timely intervention, and improve the quality of life for epilepsy patients.