PhD Thesis/Projects in FPGA Implementation of Biomedical Signal Processing Algorithms

We know that for our successful life, our health is the major issue to continue it further. So our health requires caring and monitoring system which regularly tells us about our. We need a system which analyses biomedical signals that our body generates during different activities. There are different types of analysis based on different biomedical signals such as Electrocardiogram (ECG is the electrical activity of the heart), Electroencephalogram (EEG is the electrical activity of the brain), Phonocardiogram (PCG is the ear-trumpet), Carotid pulse (CP shows variations in arterial blood pressure and volume with each heart beat), Electrooculography (EOG is used to measure the corneo-retinal standing potential and eye movements) and Speech signals etc. So to analyze these signals practically, we first need database of that signals. Here, first question is that from where we get this database? There are two approaches to get database. 1st is: get the database from different hospitals and 2nd is: some sites provide signals database free for the analyses. Here, we are telling you some online database that may be helpful for you. For the ECG, MIT-BIH database, American Health Association (AHA) and European ST-T and LTST database. There are some other databases like MIMIC, the IMPROVE and IBIS database which provides multi-modal database include signals such as activities from the brain, muscles, heart, respiration, blood pressure and others together. The PhysioNet is the free biomedical signals database website available for download. 

These entire databases contain signals with noises (it can be internal or external or combined). Here, we have two approaches to analyze these signals, one is continues and other is discrete. Discrete signals are less complex to analyze for the theoretical and practical purposes. After discretization (using Analog to Digital Converter), we need some blocks in our system that detect and remove the noises from the input signals. For the filtering and signal detection purpose, we need to know about different discrete wavelet transforms (DWT) which used scaling (for Low Pass Filter) and wavelet (for High Pass Filter) functions for filtering. We require the best architecture to implement LPF and HPF for the better efficiency of hardware. A no. of LPF and HPF are used in wavelet filter banks according to our requirement. After that main work starts that is detection of the signal with varying frequency and amplitude. Here, we need to implement in such a way that our hardware implementation is less complex, consume less power and timing performance should be better. After the processing of detection, we can compress, store, compare with particular signal, predict diseases or body condition and display the signals. All these processes in the previous sentence need different wavelet techniques, compression techniques and memories (such as registers) to store the real-time signals. This designed system should be implemented in Field Programmable Gate Array (FPGA) otherwise it is just like an algorithm (or software program which can be programmed using C and C++ also). So, to implement the system in hardware (or FPGA), we need to write the complete system code in VHDL/Verilog language stands for hardware description language used for hardware implementation.

SiliconMentor- A R&D group develops its own algorithms to minute the noise from the ECG signals and testing the same on hardware i.e. FPGA. The hardware implementation verifies the algorithms and gives an edge to take the algorithm up to the product development.

The PhD scholars and students can also approach for the research platform for their PhD research projects and PhD thesis in its core R&D domains i.e. Computer Vision, Artificial Intelligence, Machine Learning and Biomedical Signal Processing and their implementation on Hardware i.e. FPGA.

Report this Page