Project for B.Tech Students (Dr. Bharat Gupta) P-2: Broad Topic: Face recognition by robotic eye with the help of microcontroller

Transcription

1 Project for B.Tech Students (Dr. Bharat Gupta) P-1: Broad Topic: Pathological Voice Analysis Prerequisite: Understanding of Speech signal Processing, Matlab Programming, Digital Signal Processing Objective: The voice based health care/monitoring system is a non-invasive approach with various helpful aspects in tele-medical system. This system will help to differentiate between the healthy and the unhealthy individuals based on their voice biometrics. P-2: Broad Topic: Face recognition by robotic eye with the help of microcontroller Prerequisite: Knowledge of face detection algorithms, C-Programming, knowledge of Microcontroller programing and its interfacing Objective: Face recognition by the robotic eye for specific task, like - automatically attendance by facial detection. P-3: Energy Efficient MAC algorithm for Wireless Body Area Network (WBAN) Prerequisite: Knowledge of MAC algorithms, Matlab Programming, C-language Objective: Wireless systems are evolving towards the personal area networks. New very promising application are developed in the field of health monitoring using the wireless communication network near the body. These network are called the wireless body area network (WBAN), which gives the freedom to the patient in medical (health) monitoring. All the sensor in, on and around the body are battery operated, hence life time of a sensor is always an major concern. Development of MAC algorithm for WBAN in an energy efficient way

2 Project Topics: Dr. Gayadhar Pradhan 1. Speech-based emotion classification using statistical modelling based on deep learning approaches Prerequisites: Thorough knowledge of speech signal processing, MATLAB and C programming Outline of the project: Deep learning techniques are the state-of-the-art approach in the context of speech processing. Yet, not much work has been done on speech-based emotion classification. This project deals with the task of exploring and improving deep learning techniques such as DNN and CNN for speech-based emotion classification. 2. Recognizing children's speech under acoustic and ambient mismatch: Prerequisites: Thorough knowledge of speech signal processing, MATLAB and C programming Outline of the project: Recognizing children's speech is much tougher than the task of recognizing speech from adult speakers mainly due to pitch induced distortions. The problem becomes even more challenging under noisy test conditions. This project aims at improving the recognition performance simultaneously under the degrading factors. 3. Development of algorithms using deep learning approaches for speech-based bilingual keyword search Prerequisites: Thorough knowledge of speech signal processing, MATLAB and C programming Outline of the project: The keyword spotting works well for a single language domain but there is not much study in bilingual mode. Most of the conversations in Indian scenario happen in bilingual mode with English + one local language such as Hindi, Tamil, Telugu, etc. to name a few. This project deals with exploring new statistical modeling techniques and different frontend acoustic features to enhance the performance of keyword spotting system for bilingual conversation. 4. Development of speech-based person authentication system on FPGA Prerequisites: Thorough knowledge of speech signal processing, Xilinx Tools, FPGA Outline of the project: Recognizing a person using speech signal has applications in different portable devices like mobile phone. This type of application requires to work on limited amount of speech signal to avoid computational load. This projects aims to develop simplified modeling and pattern comparison algorithms on FPGA for customized development of the system. Note: Each project group should consists of at least two students and a maximum of three. Each member will be assigned a sub module on which he/she has to work independently. Depending on the permission from the funding agencies and the performance of the student, monetary incentives may be given.

3 Project List by Dr. Rakesh Ranjan, Assistant Professor, ECE department Project 1: Name of Project/Broad Title: Brief description of project work: Investigations of Cross-Talk in Multicore Optical Fiber for high capacity optical transmissions. Very soon in Future the data rate/channel capacity offered by the usual (single) core optical fiber will saturate and unable to accommodate further higher data rate demands. In view of this, Multi-core (SM & FM) fibers (MCFs) are the best candidate to address the data rate/capacity demand of the Optical transmission channel. In MCF, more than one core will be there within the same cladding diameter of 125 µm. Due to presence of more than one cores in the same cladding area, coupling (core/mode) between different cores will cause the cross-talk (XT) in the optical transmissions. MCF Technology is a very new concept for the capacity enhancement in the existing optical fiber communication systems. Objective(s): i. Study and analysis of Coupled Mode Theory and Coupled Power theory (CMT and CPT). ii. Analytical investigation for the conditions (causes) for the crosstalk between the different cores; iii. Estimation of XT between the different cores under the different core patterns/arrangements, etc. iv. Study the impact of XT on the channel capacity. v. Study and proposal for new/modified XT suppression technique(s). Pre-requisites Optical Fiber Communication, Engineering Mathematics, Electronics Devices and Circuits. Project 2: Name of Project/Broad Title: Brief description of project work: Modeling and simulation of All-optical photonic circuits for digital logic operations and Sensor applications. All optical signal processing is a very important technology and widely used in ultra-fast optical networks switching system.

4 These types of network requires very fast processes of various operations, like, Multiplexing, De-multiplexing, Bit sequence/pattern identification, switching, wavelength conversion, etc. The All-optical digital logic devices may be implemented either by using Mach Zehnder interferometer-mzi (Electro-Optic effect) or by using the Photonic Crystal-PhC (Photonic Band-gap phenomena). For MZI, the Beam propagation methods (BPM), and for PhC, the Finite Difference Time Domain (FDTD) approach may be used. These All-optical photonic circuits can efficiently utilized for the design of extended devices for the sensor applications. The expected sensor applications: Bend/twist, strain, pressure, Temperature, vibration, Gas, Humidity, etc. Objective(s): i. Study, modeling and simulation of the MZI/PhC structure. ii. Study of existing all-optical logic gates and logic devices based on MZI and/or PhC. iii. Design of a new/improved digital Logic circuits, such as reversible logic gates, combinational/sequential logic circuits, Analog comparator, etc. - Its modeling, simulations and performance evaluation/comparison with the existing logic devices/circuits. iv. Design and analysis of some of the important sensors (as mentioned above) based on MZI/PCF. Pre-requisites Optical Fiber Communication, Engineering Mathematics, Electronics Devices and Circuits. Project 3: Name of Project/Broad Title: Brief description of project work: Design of Multi-Core Photonic Crystal Fiber (MC-PCF) and its applications. The internet density is growing approx. 100 times every 9-12 years. The transmission capacity of conventional Single mode Single core fiber has also been enhanced in the past several years with the help of a variety of advanced approaches, such as the expansion of the optical transmission window and increase in the spectral efficiency, etc.

5 Designing Multi-core fiber structures using the Photonic Crystals is also a new idea to enhance the limiting capacity of the single-core optical (of PhC) fibers. Also, the PhC based MCF (MC-PCF) may have the advantage of obtaining low confinement loss as well as low dispersion. Confinement losses are the major concern in PCF and we need to keep it as low as possible. The design and analysis of the MC-PCF can be carried out using the various available techniques, such as, Finite Difference Time Domain, Beam Propagation Method, Finite Element Method, etc. This type of fiber may have potential relevance in the design of high power Laser with superior performance, because of the associated nonlinearity of PCF. Objective(s): i. Study PhC-Fiber structures and light-guidance mechanisms in PCFs. ii. Study of relevant methods (FDTD, BPM, FEM, etc.) for the implementation of MC-PCF. iii. Design of various MC-PCF structures with 'N' number of cores to achieve the condition of low confinement loss and dispersion. iv. Analysis of their respective coupling characteristics, modeprofiles, confinement losses, dispersion. v. Study of behavior of MC-PCF, by changing the fiber parameters/structures of core and/or clad regions. vi. The proposed/designed MC-PCF may be used for the design of various other photonic devices/circuits, such as splitters, etc. Pre-requisites Optical Fiber Communication, Engineering Mathematics, Electronics Devices and Circuits.

6 Project List by Dr. Priyanka Mandal List of Projects: 1. Design of a FET subharmonic mixer using ADS from Keysight Students will learn the simulation software ADS from Keysight and design a FET subharmonic mixer using it as per given specification. Objectives: Students will have sound knowledge on mixer design. Students will be familiar with the circuit implementation of a wireless transmitter/receiver system. Learn the software ADS from Keysight. The acquired knowledge will be useful for future circuit design. Prerequisites: Students should have knowledge on basic electronics and preliminary knowledge on RF and Microwave Engineering. 2. Design of a linear microstrip patch antenna array with reduced sidelobe level Students will learn the simulation software HFSS from ANSYS and design a linear array for a given sidelobe level. Objectives: Students will learn antenna array design for a given specification. Students will be familiar with various types of antennas as per the requirements of the wireless systems. The knowledge will be useful for further front end component design. Prerequisites: Students should have knowledge on RF and Microwave Engineering. 3. Efficiency improvement of a high power amplifier Students will work on class F amplifier for efficiency improvement of a high power amplifier. Simulations will be performed using ADS from Keysight. Objectives: Students will get in depth knowledge on high power amplifier design. Students will be familiar with state-of-the art research works in this area. The acquired knowledge will be useful for further circuit design using ADS. Prerequisites: Students should have knowledge on basic electronics and preliminary knowledge on RF and Microwave Engineering.

7 Dr. Seemanti Saha 1. Title: FPGA based Implementation of Baseband MIMO-OFDM Transceiver Using VHDL Prerequisite: Digital Communication, Wireless Communication, MATLAB/C Coding Abstract: The combination of multiple-input multiple-output (MIMO) signal processing with orthogonal frequency division multiplexing (OFDM) is regarded as a promising solution for enhancing the data rates of nextgeneration wireless communication systems operating in frequency-selective fading environments. This project aims FPGA based implementation of basedband MIMO-OFDM transceiver using VHDL. First, the students have to implement the basic baseband OFDM system in FPGA platform and its performance will be evaluated with both simulations and experimental test results. Next, as a test case, the OFDM-based wireless local area network (WLAN) standard IEEE a will be considered, and 2X2 MIMO-OFDM system will be implemented. This project also aims to implement the 4X4 MIMO-OFDM systems using LTE release-10 parameters. The performance of the testbed will be compared with the simulated and experimental test results. 2. Title: Digital Domain Self-interference Cancellation for Cognitive Full-duplex Communication in 5G Prerequisite: Digital Communication, Wireless Communication, MATLAB/C Coding Abstract: The rapid growth of demand for ever-increasing data rate and wireless services results in spectrum scarcity. Cognitive radio technology is used as solution for this problem. Spectrum sensing is the key mechanism used to share the spectrum in cognitive radio which will solve the spectrum scarcity problem. Conventional cognitive radio is operated in half-duplex mode which fist sense the spectrum hole and transmit if spectrum hole is found. This listen before talk approach has the constraint of losing transmission opportunities during sensing or has the Risk of undetectable collisions during transmission. This problem can be rectified by operating the cognitive radio in full duplex mode. Modern fullduplex wireless radio transceivers will be capable of simultaneous spectrum sensing and transmission. But one of the main issues in the full duplex simultaneous sensing and transmission is the residual self-interference. This project aims to implement a listen and talk algorithm to enable full duplex cognitive radio communication with the digital domain self-interference cancelling technique. With FD radios equipped at the secondary users (SUs), SUs can simultaneously sense and access the vacant spectrum, and thus, significantly improve sensing performances and meanwhile increase data transmission efficiency.

8 3. Title: Resource Optimization and Interference Management for Green Ultra-dense 5G Cognitive Radio Network (CRN) Prerequisite: Digital Communication, Wireless Communication, MATLAB/C Coding Abstract: In this project, machine learning based algorithms will be proposed that will accommodate the CR technology to maximize user throughput and network capacity of a 5G ultra-dense het-net while constraining the communication interference. Such networks will need to be highly adaptive to deal with hot spots of ultra-high device density, along with a need to be energy efficient. For network modeling, multiple network tires will be considered and randomness will be introduced in the network topology. Stochastic geometric tools will be used to evaluate the outage probability, SINR, network capacity, etc. The purpose of the project will be to show how the resource optimization and topology management could be better controlled by applying machine learning strategies, such as evolutionary algorithms and fuzzy logic, both on a local and system level. Simulation studies and analysis will be carried out to assess the performance considering both the indoor and outdoor radio channels keeping the highly heterogeneous nature of the prospective 5G network. MATLAB programming is likely to be the simulation tools. The objective of the project is to investigate how cognitive radio techniques can be applied in ultra-dense small cell 5G network to enhance its performance. It will also investigate how backhaul diversity can be achieved in the network. The benefits and drawbacks of such the proposed approach compared with more conventional strategies will be evaluated. The deliverable of the project is adaptive and energy efficient network topology management and radio resource optimization schemes for 5G small cell networks.

9 Project List by Dr. Moina Ajmeri Proposed Projects: 1. Comparative study of various reported works to recommend suitable tuning method for the given process model. A large number controller design methods are reported in the literature for stable, unstable and integrating etc process models. These methods differ from each others in terms of design method involved, control structure or controller configuration used. A comparative study of those reported works may be done to help the user in selecting suitable method for his/her plant. 2. To propose new controller design methods for stable and unstable process models. Transfer functions of the process are assumed to be known to us. We have to select a control structure/block diagram and the controller configuration (P/PI/PID etc). There are various methods for designing/tuning controller. We need to choose one of these method to find parameters of the selected controller. In order to justify our method we need to compare it with the previously reported works 3. To propose an Smith predictor control scheme (It gives improved performance for processes with large time delay) for integrating plus time delay process models. Here we need to assume integrating transfer functions with some delay. Many control structures are reported by various authors. We have to take one of them and then using some technique we can propose new tuning rules. Prerequisite: MATALAB coding and basics of control theory

10 Project List by Dr. Gaurav Kaushal Project 1: Name of Project/Broad Title: Analysis of radiation hard SRAM cell using FPGA board Objective(s): vi. Implement the SRAM cell into FPGA board using HDL language vii. Make an interface between FPGA board and matlab/function generator viii. Apply the signal from matlab/function generator to FPGA board ix. Analyze the SRAM array performance Pre-requisites Basic knowledge of MATAB, Verilog language, and SRAM circuits Project 2: Name of Project/Broad Title: Implementation of 8 8 SRAM array using h-spice Objective(s): v. Implement the SRAM cell, decoder and sense amplifier in h- spice vi. Make the interface between decoder, SRAM cell and sense amplifier vii. Analyze the SRAM array performance Pre-requisites Basic knowledge of digital circuit, spice software