Miscellaneous

– Issues which the user prefers to be classified by the moderator or thinks lie outside the purview of above four segments.

Packet Loss Concealment and Quality Assessment for Voice over Internet
7arjuna@iimahd.ernet.in

Today’s internet offers best-effort service due to its nature, and speech quality in transmission over internet networks suffers mainly due to packet loss, delay and delay jitter. Playout buffer at the receiving side can be used to compensate for the effects of jitter based on a tradeoff between delay and loss. Our objective here is to investigate the effect of packet loss on speech quality in voice over internet scenarios. Assessing speech quality is inherently subjective issue. However, subjective voice quality evaluation is impractical and there are objective methods to fulfill the task of speech quality assessment [1][2][3]. We look at the E-Model and observe the “MOS Vs Loss” and “R Vs Loss” trends for iLBC and the ITU-T G.711 packet loss concealment algorithm applied on both Male and Female utterances in English and Hindi.

DESIGN AND OPTIMIZATION OF LOW POWER CURRENT OP-AMP (COA)
sourdon@gmail.com
nipunrastogi@gmail.com

Signal manipulation in current mode, has been of a lot of interest to researchers for quite
some time now. This attention from the researchers can be attributed to the distinct
advantages offered by current mode circuits, and also because of the sensor applications
where, in many situations, signal conditioning in current mode is the only viable option.
The circuits for signal processing in current mode have continued to evolve with the
changing demands, performance constraints, and technology scaling. The current
operational amplifiers (COA), as the most fundamental signal manipulation blocks have
also gone through many generations of design. Over the years the design has changed and
evolved to achieve many improvements like higher gain and bandwidth, improved
robustness towards noise and process variation etc. As the constraints on power
dissipation get more stringent and the need for high performance in terms of gain,
bandwidth, and robustness persists; the scope for improvement in the latest COA design
is as good as ever. This project focuses on the design of a novel wideband current op-
amp, suitable for low power applications. The various topologies available today exploit
one tradeoff or other, to achieve higher bandwidth, while trying to preserve the other
performance parameters. We begin by exploring different topologies available and their
respective tradeoffs. We propose some modifications in the selected topology, and
employ voltage and technology scaling, in order to lower the power requirements of the
circuit, without compromising the performance of the circuit. In order to optimize the
circuit computationally, a mathematical formulation of performance parameters of the
circuit was done. For optimization computational, a basic approach and an approach
based on genetic algorithms were employed and compared. Extensive simulations were
done to analyze the performance of the circuit, and to compare it with other topologies
reported in literature.[u]

Content-Based Video Retrieval
tanysachan@gmail.com

Recent advances in computing, communication and data storage have lead to an increasing
number of large digital libraries that are becoming publicly available via the Internet. The current
systems become insufficient day-by-day as the archives grow very fast and become huge. They
are insufficient because they search the data by queries from the archives that are prepared
manually by attaching a text value for annotations of a video entry while putting it to the archive.
Then the system searches the queried information from these annotations and matches the
queried information with the videos that have correct annotations and only the exact matching
word is being searched within the annotations.
Huge archives made this process inefficient. However, our goal is to make the video archives
available to a great variety of users in an efficient, effective and organized way. It is basically
content based video retrieval (CBVR) system. After addressing basic concepts and techniques,
we would be concentrating on the semantic gap problem, i.e. the problem of inferring semantics
from raw video data, as the main problem of CBVR.
We will be addressing the problem of tracking tennis players using image processing algorithms
from a tennis video sequence. After analyzing all the collected data by means of geometric,
mathematic and physical procedures; all kind of strokes will be detected. The detection of stroke
sequences during a whole set will be our objective. Further we will demonstrate the validity of a
stroke detection algorithm. This characteristic has potential applications for automatic editing,
broadcasting, archiving, browsing and training.
Our algorithm is based on finding the feet position of the tennis players and their acceleration
based on the fact that before a stroke there is almost always a short run. To cope up with the
limitations of using acceleration based approach, a racquet and ball detector along with audio
system should be implemented but this goes beyond this B.Tech Project.

To design a RFID based inventory control system

rohitkaran@gmail.com
arpit.srivastava.iitk@gmail.com


The properties of RFID vary dramatically with the surface on which it is mounted. This poses serious threat to the use of RFID in inventory control, limiting its advantage over the traditional barcodes. The biggest advantage of the RFID tag, to be detected without being in close proximity to the antenna, is lost when it is attached to a conducting surface.
Hence, in this project we have studied the varying range of RFID, when attached to different insulating and conducting surfaces, in order to find out how the effect of the conducting surface can be nullified.
Another aspect of the project was to develop a series of Labview programs for communication between the reader and tag. Programs gathering and setting information like power value, protocol, Tag Information, etc. were made which could be further used for Inventory management in Laboratories and libraries.

Designing a voice controlled robot
raman1985@gmail.com

Our project aims at building a robot, presently, voice controlled, which accepts
inputs from the programmed computer and performs five basic mechanical actions of
forwarding, reversing, turning left and right and stopping. We chose Speech control
because in future it is expected that speech recognition systems will be used as the main
man-machine interface for robots used in rehabilitation, entertainment etc.
This project report discusses in detail the proposal of the project, used methodology,
experiments carried out and their results. It also discusses the future work to be carried
out as part of the major project and the strategies to implement and achieve them.

vijaykuppa@gmail.com

Particles and structures with sizes in the nanometer range (10-9 m) are known as nanoparticles and nanostructures respectively. In recent times nanoparticles and nanostructures are of great importance as these are the raw materials for various industries like semi conductors, microscopic instruments, drug delivery etc. The uniqueness of nanoparticles comes from the fact that they lie between bulk and atomic dimensions [1] and are therefore endowed with special properties, thereby making them starting material for many futuristic applications. These properties are attributed to the scale of the size of those materials which is of the order of a few hundreds of molecules.
Nanoparticles with controlled properties are of scientific as well as technological interest [2] due to the increasing number of promising applications for such materials and the technical-scale production of nanoparticles with tailored properties is of particular interest due to the increasing demand from industry. The properties of nanoparticles depend on the crystallite sizes and many properties change drastically if crystallite sizes fall below the submicron range. Particles with a size between 1 and 100 nm have the special benefit that physical and chemical properties, e.g., melting point, solubility, surface hardness, electrical conductivity and surface reactivity, differ from the bulk material properties and are strongly size dependent. The mean particle size and the particle morphology, but also the modality and the broadness of the particle size distribution (PSD) are important indications for the quality of a particulate product. By controlling these characteristics during the synthesis it is possible to produce advanced nanomaterials with tailored properties. Therefore, control over the particle sizes and particle size distributions (PSDs) is highly desirable.
Nanoparticles have a high specific surface area, inter-facially dominated properties and electrical/optical properties that can be distinguished with respect to corresponding bulk materials on the one hand and individual molecules on the other [3]. Therefore, nanoparticles find applications in diverse fields such as heterogeneous catalysis, microelectronic devices, information storage, optical devices, pharmaceuticals, paints, ceramics, semiconductors, advanced biomedical applications, and nanotechnology.
To meet the primary specifications of these applications, a precise control over the particle sizes and particle size distributions (PSDs) is necessary, during their preparation. But, even though nanomaterials are becoming increasingly important in various technological applications, the large-scale production of tailor-made nanoparticles still represents a particular challenge. Different engineering approaches such as gas-phase synthesis or bulk phase precipitation in liquids show several disadvantages, as the non-ideal mixing in these processes often leads to broad particle size distributions and an average particle size which is difficult to control.

Medical Assistance In A Distributed Environment Using Web Agents and Artificial Intelligence
nishithg@iimahd.ernet.in

The project Medical Assistance in a Distributed Environment using Artificial Intelligence
and Web Agents is based on the concept of telemedical systems which would
integrate the mobile agent technology along with the applications of Artificial Intelligence.
Such systems are expected to perform in a heterogeneous environment.
This system would act as a facilitator to the medical professionals for the quicker
diagnosis and providing efficient treatment.

chinmayakv@iimahd.ernet.in

Individual (single) component pervaporation study helped to address some of the basic curiosities for the process of pervaporation. Investigations were carried out to focus on the location of vaporization during single component pervaporation. A mathematical model was developed for single component permeation during pervaporation, assuming two zones inside the membrane; namely, liquid permeation and vapour permeation zones. Considering a pressure distribution across the thickness of the membrane, Kelvin equation (saturation vapour pressure gets modified inside the membrane due to permeant membrane interactions) proved to be useful in developing the model. According to the model assumptions, the sorbed liquid first transports as liquid; and as soon as it finds the region, where pressure is Kelvin pressure, it evaporates and continues to transport as vapor. Further, the developed model was found to be useful in describing the flux in terms of downstream pressure variations. Accordingly, location of vaporization was determined. It was observed that vapor phase transport dominates in the membrane at low downstream pressures. Importance of consideration for both the phases, during modeling, is discussed. Activity profile, determined across the membrane, was observed to be in agreement with the experimental observations (as per literature). The study may help to establish a fundamental framework in turn to model for binary and/or multi-component mixtures.

Computational Methods for Aerodynamics of Wind Turbines
shantanu.iitkgp@gmail.com

Horizontal axis wind turbines can experience significant time varying aerodynamic loads, potentially causing adverse effects on structures, mechanical components, and power production. The needs for computational procedures for investigating loading and dynamic response have increased as wind turbines become lighter and more flexible.
The present works deal with the development of an aerodynamic simulation has been done in two parts:
a) Simulation of horizontal axis wind turbines for computation of loads on wind turbine blades through MATLAB. Codes were written in such a way that it should calculate loading according to different wind profiles that is by taking in account for wind shear(due to atmosphere),wind turbulence and tower shadow.
b) The second part of this project is to develop a procedure to numerically model airflow over airfoils using GAMBIT and FLUENT. Two dimensional model for the airfoil S809 was created, drawn and meshed in Gambit using geometry data gathered by the National Advisory Committee for Aeronautics (NACA). The model was read into FLUENT where boundary conditions flows were applied and the
discretized Navier-Stokes equations were solved numerically. The airfoil section lift and drag coefficient from the numeric simulation was compared with experimental data from the literature and shown to agree within 10% for angles of attack below 10°.

VLSI Architecture for one dimensional media filtering
teja.iitkgp@gmail.com

This work presents two VLSI architectures for one dimensional median filtering, to detect and suppress the impulse noise in real time signal processing applications. The first proposed architecture is a fast VLSI architecture for nine samples window, and the second one is an extended and modified form of the first architecture know as flexible architecture. The flexible architecture can be used for different window sizes ( i.e., 3,5,7,9) as well as different sample word size (i.e., 8-bits and 16-bits) in a single chip with some extra hardware for controller circuitry. Both the architectures have been synthesized with commercially available 0.18µm CMOS technology. The synthesis results show that the fast architecture and the flexible architecture lead to throughputs of 700MSamples/sec and 600MSamples/sec respectively

VIDEO COMPRESSION ARCHITECHTURE
teja.iitkgp@gmail.com

This work presents the design of a novel Video Compression Architecture employing DCT. A custom made Radix-4 Cordic Processor has been designed and used in the DCT Architecture to accomplish greater throughput and thereby greater framerate. In order to have a clear picture of the advantages of the custom made cordic architecture, two more architectures have been implemented in Verilog HDL and the outputs compared. Their throughputs are 250MHz and 300MHz, while our architecture scales upto 450MHz, thereby making it an ideal choice for Real Time Video Processing applications.

MICROBIAL FUEL CELLS- convert biomass into energy
btushar02@gmail.com

Microbial Fuel Cells (MFC) are bielectrochemical transducers that convert biomass into electricity. The present paper reports the studies of a two-chambered MFC, using Enterobacter cloacae IIT-BT 08 in MYG medium. Two MFCs namely, dialysis membrane MFC and Salt Bridge MFC were constructed for studies and various parameters were studies and optimized. The effect of different electron mediators, concentration of the mediator, ionic strength (salt concentration) of the medium, surface area of the salt-bridge and the size of the electrodes on the power generation in MFCs are reported. In case of methyl viologen (MV) as the electron mediator, the voltage generation was 0.4 V but no current was detected. Different concentrations of methylene blue (MB) were also studied as the mediator. Maximum voltage of 0.37 V was seen at 0.05 mM MB whereas a maximum current and power of 56.7 µA and 19.2 µW respectively were observed in case of 0.03 mM MB. Corresponding power density and current density of 9.3 mW/m2¬¬ and 27.6 mA/m¬¬¬2¬¬ respectively were obtained. When the surface area in contact with the anode and cathode chambers was increased, a proportionate improvement in the power output from 19.2 µW to 708 µW was detected. Maximum power density and current density of 236 mW/m2 and 666.7 mA/m2 respectively were observed which are found to be very promising for a salt bridge MFC. When the surface area of the electrode was increased step wise, it showed an initial increase in current density as well as power density but the same decreased for large surface areas of electrodes, indicating that the surface area of electrodes is dependant upon the cross sectional area of the salt bridge used. A new electrode, Platinum coated Tantalum, has been studied as a cathode electrode.

DIGITAL SIGNAL PROCESSING, implementation on a DSP processor
rinky3011@gmail.com

The main goal of this thesis work is to learn and understand the features of image processing and afterwards, once familiarised with this standard, to develop a code for image processing and implementation on the DSP processor. Matlab/Simulink is used for designing the model and Real-Time Workshop and Code Composer Studio for targeting that Simulink model on the DSP board in order to have a better understanding of the standard and of the system performance running this model under both platforms
One of the purposes of the thesis, to target the whole simulink model on the dsp board, could not be achieved. The cause of this is not clear; it could be due to the fact that the software used was not the suitable one. Another option could be that there was a wrong use of such software. A third possibility is that there was a problem with the configuration.

For future works we are trying with Embedded Target, using the current version. One hint for future works is to try to implement an image processing and target this on the DSP using the software Embedded Target for c6000. Some libraries can be downloaded from the TI page (see references at the end of this work) with very useful functions that makes code generation much easier.

The whole model in the dsp could not be run but one important block in whole project, the, codes of simple nature was implemented and run in the DSP processor and in the CPU. By achieving this, the following interesting conclusions are reached. As it was commented in the introduction about digital signal processors, that kind of processors have some advantages compared to normal microprocessors in terms of speed or memory usage. Due to the features of the DSP processor, they are quite effective in implementations where mathematical operations are involved, for example in a wireless system. As we could see in the results obtained in the analysis of the system performance, the DSP processors are much faster than a normal microprocessor and require less memory usage. For this reason the use of DSP processors is highly recommended for this kind of implementations. The DSP offers a large number of possibilities for improving the performance of the implementation as well, like debugging, memory allocation or synchronization techniques. We can conclude that if in the FFT implementation the best results are obtained by using DSP the same can be expected in a whole transmission wireless system.

HEAT PUMP DEHUMIDIFIER
prabhats@gmail.com

Drying is one of the oldest and most cost-effective means of preservation of grains, crops and foods of all varieties. Out of the various drying processes viz. conventional drying, vacuum drying, microwave drying, freeze drying, sonic drying, osmotic drying, drying with one drying system in combination with other to enhance the product quality as well as to reduce the energy requirement, conventional drying method is the most commonly used. But this drying requires large quantities of energy for heating and water removal, and a relatively high temperature moist air is allowed to leave from the dryer, which is not only considered as a waste but also detrimental to the environment. To overcome these drawbacks of conventional drying system, heat pump is added with the conventional system, called heat pump dehumidifier A HPD can be operated over a wide range of temperatures, providing good conditions for the drying of heat sensitive materials. The technology requires less energy, as the system can recover the latent heat in a closed loop, and can be independent of ambient weather conditions. In the present Study 1.0, 1.5 and 2.0 TR capacity heat pump is used for drying during 6 hours interval keeping the RH of atmospheric air 70% as constant and then in is found out that at what combination of mass of product to be dried and Mixing of fresh air the power requirement of to operate the heat pump is minimum.

SILICON CARBIDE- A NEW WAY
sireeshbabusukhavasi@gmail.com

Silicon Carbide was produced by Chemical Vapor Deposition of borosilicate glass in a Graphite furnace at a temperature of 1300oC. First phase of the investigation was carried out with the preparation of SiO2 sample using borosilicate glass. It was found that formation of short silicon carbide whiskers had taken on the surface of the graphite walls. Whiskers of about 600 μm were observed at 1300oC at a pressure of Mpa.
From the present investigation it has been concluded that:
a) Silicon carbide can be produced using borosilicate glass.
b) Silicon Carbide Whiskers produced using borosilicate have better mechanical properties than produced by other conventional techniques due to the presence of B4C

WATER LUBRICATED TRANSPORTATION

Water lubricated transportation is a method of transporting highly viscous fluids at low cost. A viscous fluid forms a core surrounded and lubricated by a water annulus. The water reduces the shear stress on the wall of the pipe. So, the core doesn’t feel the effect of the wall. And the frictional pressure drop is just in the order of the less viscous fluid. An important series of experiments on the water lubricated pipelining was carried out by Russell and Charles (1959); Russell et al. (1959), Charles (1963) and particularly by Charles et al. (1961). Glass (1961) and others found that the lowest pressure gradient was achieved when the input ratio of water to
oil was between 30% and 40%. Other experiments on water lubricated transportation in horizontal pipes were reported by Stein (1978) and Oliemans et al. (1985). Bentwich (1964) studied for the laminar case with the basic core–annular model. He solved the Poisson equation that governs the eccentric core flow model with a Fourier series in bipolar coordinates. However, he did not use his solution to evaluate the friction factor or hold-up ratio. Huang et al. (1994) studied laminar and turbulent core–annular flow to assess the effects of eccentricity and the volume ratio on the friction factor and holdup ratio. They used a model of core–annular flow in which the oil core is a perfect cylinder with generators parallel to the pipe wall, but off-center and adopted a standard k–ε model with a low Reynolds number capability for turbulent case.

Implementing Authentication and Authorization in NSIS Protocol Suite

pragup@gmail.com
ding.iitk@gmail.com

Due to evolution of Web 2.0 and increased usage of multimedia services, the in-
ternet infrastructure has seen a phenomenal increase in real time data traffic. This calls for an
extensible and secure signaling protocol. NSIS (Next Steps in Signaling) [2] [3] is a major initiative
by the Internet Engineering Task Force(IETF) in this direction. It is a generic IP based signaling
protocol which provides means to establish and manage network control states along a data path
between two nodes communicating on the Internet. The modification of control states along a
data path allowed by NSIS introduces the need for authentication and authorization. This report
discusses our work done in CS498 and CS499 leading to the implementation of authentication and
authorization object in NSIS.