Indian Institute of Science Bangalore
etd@IISc Electronic Theses and Dissertations at Indian Institute of ScienceNot a member yet
6204 research outputs found
Sort by
Relating Representations in Deep Learning and the Brain
No full textDeep Neural Networks (DNN) inspired by the human brain have redefined the state-of-the-art performance in AI during the past decade. Much of the research is still trying to understand and explain the function of these networks. In this thesis, we leverage knowledge from the neuroscience literature to evaluate the representations learned in state-of-the-art language models. We use sentences with simple syntax and semantics (e.g., “The bone was eaten by the dog.”), and train multiple neural networks to predict the part of speech, next word. We present other sentences of this same simple form, word-by-word to humans in a magnetoencephalography (MEG) scanner for silent reading and comprehension. We then train a linear regression model to predict observed brain recording from the hidden layers of the trained neural networks and popular pre-trained networks like BERT and ELMo.
We find that the middle layers of these networks are the most predictive of the recorded brain activity. But, a more fine-grained evaluation shows that various types of stimuli (determiner, adjective, noun, verb) are represented more dominantly in different layers of the language model. Further, we test the semantic composition capabilities of these networks with respect to the human brain. Semantic composition is defined as the rule-based combination of the parts that constitutes the meaning of the whole. We collect new data and develop a new framework to perform this evaluation incrementally as each word in the sentence is processed in the brain and DNN. As a result, we are able to analyze the effect of the composition function in representing the same word as more of the sentence context becomes available. Our experiments show that DNN models are effective in encoding the sentence being read and are able to predict the word which occurred earlier in the sentence, indicating good composition. We find that in these tests, the right frontal and right temporal brain regions are predicted with best accuracy. Previous research has suggested that these brain regions are responsible for executive and memory function.
As an additional contribution, we propose a new dynamic time warping based distance metric to evaluate alignment between the predicted brain activity versus the observed brain activity. The new metric helps tackle the variability observed in a single subject’s recorded brain activity.Ministry of Human Resource and development India (MHRD), Pratiksha Trust, and CMU BrainHu
Biochemical and structural studies to provide insights into initiator tRNA delivery by eIF2A in noncanonical translation initiation
No full textEukaryotic initiation factor 2A (eIF2A), a non-canonical translation initiation factor, delivers initiator tRNA (tRNAi) onto the 40S ribosomal subunit in a codon-dependent and GTP-independent manner, unlike the canonical initiation factor eIF2. eIF2A helps to perform the stress response under global protein synthesis inhibition by translating some specific mRNAs, hence regulating their protein synthesis. However, the underlying mechanism of this eIF2A-mediated non-canonical translation initiation is not known. It has a biologically significant role in various diseases like early tumorigenesis, neuromuscular disorders, viral infection, and Integrated stress response. Hence, it is important to elucidate the role of eIF2A in translation initiation. We aim to understand the detailed mechanism of eIF2A-mediated translation initiation by biochemical and structural approaches.
In this study, I have characterized the yeast eIF2A protein in-silico, biochemically and structurally. The computational analysis of eIF2A sequences from various organisms revealed unique features in eIF2A, namely nine-bladed beta-propeller domain, a conserved motif, stretch of positive residues, and C-terminal helices. It also revealed their diversity and presence of different isoforms of eIF2A in various organisms. Biochemically, I purified recombinant yeast eIF2A protein and detected interaction with 40S ribosomal subunit and initiator tRNA. Mutational studies were done to figure out the 40S and tRNA binding regions in eIF2A. eIF2A uses its unstructured middle regions and C-terminal helices for binding tRNA and 40S, respectively. The exact role of conserved beta-propellor of eIF2A remains to be deciphered.
Further, reconstitution of eIF2A complexes with 40S was used for structural studies by cryoEM. Multiple maps were reconstructed for different ribosomal complexes with eIF2A. These maps contain extra density, tentatively assigned to eIF2A, at the subunit interface of the head region of 40S. The ribosomal proteins present in the 40S head indeed bind with eIF2A in pull-down assays providing support to the assignment of extra density to eIF2A. This position of eIF2A is near the P site adjacent to tRNA and explains how eIF2A may deliver tRNA. Further, this position would have steric clash with the alpha subunit of the canonical eIF2 complex but not with any other eIFs. The insights helped us propose the mode of tRNA-delivery by eIF2A in translation initiation
Radical and Lewis Acid Mediated Syntheses of Panaginsene, Indole-Fused Azabicyclo[3.3.1]Nonane and Davanoids
No full textThe thesis contains radical and Lewis acid-mediated syntheses of panaginsene, indole-fused
azabicyclo[3.3.1]nonanes, and davanoids. The thesis includes three chapters; chapter-I
contains the total synthesis of panaginsene, which is an angularly fused tricyclic molecule
having a quaternary carbon center. A Ti(III)-mediated radical cyclization protocol was used to
construct the C-C bond, which yielded the quaternary carbon center. The tetra substituted olefin
was synthesized by McMurry olefination reaction. Chapter II focused on constructing an
indole-fused azabicyclo[3.3.1]nonane core of many alstonia alkaloids, which was synthesized
using Sm(II)-mediated radical cyclization reaction. The precursor tetrahydro β-carboline was
synthesized by the Pictet-Spengler reaction. A novel radical cyclization protocol was
developed to achieve the azabicyclo[3.3.1]nonane core annulated with an indole ring. Chapter
III contains a Lewis acid-mediated cycloetherification reaction to construct the tetrahydrofuran
core of davanoid natural products. The side chain of davanoids was introduced by Grignard
addition on Weinreb amides of davana acid. The non-Evans syn aldol reaction was used to
achieve the enantioselective synthesis of the chiral centers of davanoids. Some of the key
reactions utilized in this thesis are Ti(III)-mediated radical cyclization, Michael reaction,
Sharpless asymmetric epoxidation reaction, HWE reaction, McMurry olefination reaction,
Pictet-Spengler reaction, Sm(II)-mediated radical cyclization, Wittig olefination reaction, non Evans syn aldol reaction, Lewis acid-mediated cycloetherification reaction, and Grignard
reaction
Space-Time Gauge Theories for Continuum Modelling of Viscoplasticity, Damage And Electro-Magneto-Mechanical Phenomena in Solids
No full textOver the years, sustained research efforts have aimed to understand the material behaviour
under a broad range of response regimes, especially from micromechanical or phenomenological
perspectives — via both continuum modeling and experiments conducted at different
scales. However, a review of the relevant literature has revealed that physics-based models
that can replicate experimental results are very few, and models depicting consistent coupling
phenomena observed in solids beyond elasticity are elusive. Symmetry-driven approaches to
continuum mechanics of solids typically have a unifying nature, combining the prediction
of diverse observed phenomena under a single umbrella. This thesis attempts to derive a
unified field theory for various physical phenomena in solids by exploring local symmetry,
which offers a framework to consistently arrive at the relations among polarization vector,
temperature, scalar potential, vector potential, and the electric and magnetic field for multiphysics
phenomena. Furthermore, this approach enables a consistent and robust coupling
among flow stress, strain rate, and other variables describing the kinematics of plasticity and
damage.
This thesis draws upon continuous and local symmetry-based principles of gauge theory
to arrive at continuum models for various electro-magneto-mechanical coupling phenomena
and inelastic responses involving plasticity and damage in solids. The specific local symmetries
we exploit in the process are conformal (scaling) and translational in space-time.
The work presented may thus be classed in two parts – one focusing on a unified continuum
description of multi-physics phenomena such as piezoelectricity, piezo-magnetism, coupled
thermoelasticity and flexoelectricity and the other on dissipative phenomena such as plasticity
and damage.
Under an inhomogeneous (local) action of the symmetry (gauge) group, invariance of the
energy density is lost. Minimal replacement is used to restore gauge invariance of the energy
density; this requires the definition of a gauge covariant operator in place of the ordinary
partial derivative. Minimal replacement introduces a non-trivial gauge compensating 1-form
field. The 1-form field is decomposed into an anti-exact part and the exact differential of a
scalar-valued function. The other essential ingredient of gauge theory is minimal couplin
Representation of Natural Stimuli in Neural Signals across Scales and Frequencies
No full textNeural activity from the brain can be recorded at different scales using a variety of electrodes, which vary in their resolution, cortical spread and invasiveness. The electroencephalogram (EEG) is recorded from the scalp, electrocorticogram (ECoG) from the cortical surface, while microelectrodes are inserted into the cortex which record local field potentials (LFP) and spiking activity in animals. These signals have been used to drive Brain Machine Interfaces with varying degrees of success, but an objective comparison of their efficacy has not been performed. A sensory system such as the visual cortex can be used as a model to compare the information available across these scales.
In this work, using a customized hybrid array containing both micro and ECoG electrodes, we recorded simultaneous signals from up to four scales (spikes, LFP, ECoG and EEG) from the visual cortex of two monkeys while they viewed a large array of natural images as well as parametric stimuli such as gratings. Complementary information theoretic and decoding approaches were used to quantify the information content about naturalistic and parametric stimuli at each of the scales. We found that the information content in ECoG exceeded all other measures, including spiking activity. Further, the maximum information content was found in the gamma (30-80 Hz) frequency range of the signals.
Several theories have been proposed to explain a potential role of gamma oscillations in the coding and visual information and its communication across brain areas. We instead tested whether gamma oscillations elicited by natural images could be explained simply based on the local image properties. To do this, first the gamma response for multiple visual features (such as orientation, spatial frequency, size, contrast, hue, saturation etc.) needs to be determined. Though the dependence of gamma on such features has been well studied when presented alone, how these features jointly affect gamma has not been investigated in detail. We found that gamma responses to a pair of features were largely separable in both LFP and ECoG. Based on this, we developed a multiplicative model in which the response to multiple features is simply a scaled product of individual features, and used it to predict the gamma responses to parametric gratings and chromatic patches.
Finally, we built an image computable model to predict gamma responses to complex natural images by extracting simple features from them and incorporating the previously learnt dependencies of gamma response. Our model was able to estimate the gamma responses to both chromatic and grayscale images.
Overall, the comparative study of information across scales can help in designing more accurate and reliable BMIs, while the predictability of responses can be used to increase the precision of BMIs. The prediction of gamma responses based on low level features also offers a simple “null” model based on local image properties, against which more advanced theories of gamma based on predictive coding or selective communication can be tested
Lab to Point-of-Need Technology: Solving the A3 puzzle of in-vitro diagnostics
No full textThe role of medical diagnostics is morphing rapidly in the face of imminent challenges such as pandemic preparedness. Correspondingly, in-vitro diagnostics are evolving — growing beyond
the traditional Centralized Diagnostic Laboratories (CDL) and now include Point-of-Care Diagnostic (PoCD) devices and lab-on-wheel setups (such as medical diagnostic bus/van) as well.
However, even now, the triple challenge of simultaneously meeting Accessibility, Accuracy and Affordability (A3) still remains unaddressed by and large. This implies that all the
current diagnostic paradigms either do not provide a sufficiently large range of assays (lack of accessibility), or do so at the expense of accuracy and/or affordability. Unfortunately, there
exist several scenarios where meeting this A3 criteria is critical to delivering medical diagnostics.
This thesis proposes and builds upon a diagnostic paradigm that has the potential to address this A3 challenge. It highlights how the current, diffused efforts towards miniaturizing
CDL or merely building more PoCD devices would not serve the purpose. This is because (a) existing diagnostic instruments cannot be miniaturized sufficiently to be easily transported to
points of need, and (b) the underlying technologies of existing instruments are fundamentally different and mutually incompatible, such that their operations and consumables cannot be
integrated into one unit. This thesis proposes that research and development efforts should instead be re-oriented towards building a modular, portable diagnostic platform complemented
by assay-specific, disposable, “smart” consumables. The platform would contain an integrated set of novel/innovative instrumentation that can carry out all basic unit operations needed for
in-vitro diagnostic assays. The platform’s capabilities would be complemented by the “smart” design/processing elements available on the consumables, which provide the remaining needs
for conducting the assay. This framework of segregated functionality ensures a compact and low-cost platform as the demands imposed on it are kept minimal. This framework also assures
future readiness since the consumable can be continuously re-engineered for improved performance as well as handling any new diagnostic assay as and when needed. The combination of the diagnostic platform and smart consumables is termed here as Lab to Point-of-Need Technology (LPoNT). Such technology needs novel/innovative methods for sample preparation,
bio-analyte processing, read-out, reagent storage and fluid handling to transcend the barriers imposed by the existing diagnostic methods. Emerging technologies such as microfluidics and
image processing are instrumental to such developments. This thesis presents an opto-fluidic LPoNT platform designed along these lines, and the instrument is demonstrated to perform microscopy imaging as well as optical absorption based measurements, in a semi-automated manner. The platform can switch between the two modalities in a few seconds, allowing a single sample to undergo any of the thousands of tests pertaining to these modalities (Cytology and Biochemical assays). This would otherwise need two different instruments and also different associated consumables as well as operator training. The microscopy unit achieves a resolution of 0.78 µm and the optical absorption unit can acquire broadband (300-800 nm) measurements with a high sensitivity photosensor. The developed platform is low-cost, portable enough to be carried in a backpack, and is modular, extendable.
It uses off-the-shelf optical, mechanical and electronic components integrated in a manner that provides a single user-interface (physical and digital) to all functionality.
Three methods have been devised to fabricate “smart”, low-cost, disposable consumables that support the utility of the developed platform. The first method demonstrates rapid construction of a low-cost, optically clear cartridge with a fixed volume cavity. This “smart slide” can be used for both, volumetric microscopy as well as optical absorption based measurements. The second method demonstrates “smart curing” of the polymer polydimethylsiloxane (PDMS) and reduces curing time from hours to as low as 2 minutes, without any additional
chemical or physical treatment. It is targeted at soft lithography performed with a silicon master mould, and uses a commonly available microwave oven for curing. The method could
assist commercial-scale fabrication of PDMS based in-vitro diagnostic devices, many of which are reported in research and are amenable to be used as LPoNT-compatible consumables. The
third method demonstrates single-shot, cleanroom-free “smart fabrication” of electrodes and microfluidic channel in low-cost commodity copper coated boards. The “smart” processing
element of micro-electrodes would allow operations such as cell counting and membrane lysis, which form essential sub-steps of some assays. Electrodes spaced as close as 50 µm and microchannels as narrow as 240 µm have been constructed. This method lowers the cost barrier to usage of electric-field based bio-manipulation methods in LPoNT consumables.
Experiments were conducted with the developed LPoNT platform using clinical whole blood samples, pre-prepared dried smear slides and synthetic solutions of bio-chemical analytes. Experimental results report precise measurement of blood Hemoglobin and Creatinine concentration, and generation of clinically-usable microscopy images of cells. Hemoglobin in clinical
samples has been measured with an average error of 2%, which is well below the acceptable limit of 7%. These results pave the way for a large set of assays, including complete blood cell
count (CBC) and screening for hemoglobinopathies such as Thalassemia
Deciphering the mechanisms employed by Salmonella to mediate modulation of host endo-lysosomal machinery
No full textSalmonella genus encompasses Gram-negative, rod-shaped, facultative anaerobic, non-sporulating, and predominantly motile enteric bacteria. It causes pathogenesis in a wide array of hosts, from cold-blooded animals to humans and diseases ranging from occasionally fatal systemic fever (Typhoid) to self-limiting diarrhea (gastroenteritis) in humans. Regardless of current medical advancements, Salmonella remains a significant cause of morbidity and mortality in developing countries.
Salmonella remains enclosed in a vacuole inside the host cells, and these modified vacuoles are called Salmonella-containing vacuoles (SCV). In general, intracellular pathogens are subjected to the various host defense mechanisms, among which avoiding fusion with lysosomes is a prime challenge. Several intracellular bacteria have employed strategies to escape lysosomal degradation, such as Listeria spp., Shigella spp., E. coli K12, and Mycobacterium tuberculosis. Similarly, during the intracellular life of Salmonella, the SCVs are also targets for the lysosomes. Hence, Salmonella substantially alters host endo-lysosomal pathways to thrive inside the host cell.
In our study, we have tried to decipher a novel molecular mechanism employed by Salmonella to modulate the host endo-lysosomal machinery to establish its replicative niche in otherwise hostile host cells. We found that Salmonella pathogenicity island 1 (SPI-1) effector SopB subverts host xenophagy by altering the phosphatidylinositol phosphate dynamics on the SCV membrane. It also mediates the downregulation of host lysosomal number, which is one of the prime strategies for establishing a replicative niche inside the host cell. Several of these fusion events of endo-lysosomal pathways are with the help of host syntaxin proteins. We further identified a crucial interaction of host Syntaxin 3 (STX3) with SCVs and observed that this interaction is SPI-2 dependent. Furthermore, this interaction with STX3 facilitates the division of the bacterium along with SCV and maintains a single bacterium per vacuole status, thus, facilitating a hospitable niche inside the host cells.
Part I: To unravel the mechanisms employed by Salmonella SopB to subvert host cell xenophagy in macrophages.
Salmonella survives and utilizes macrophages for effective dissemination throughout the host, causing systemic infection. One of the central host defense mechanisms in macrophages is xenophagy, or macro-autophagy where invading pathogens are targeted for degradation by fusion with lysosomes. The process of autophagy depends on the dynamics of phosphatidyl-inositol phosphates (PIPs) species present on the vesicular membranes. Salmonella, when inside intact SCV, escapes autophagy. However, the exact mechanism remains unknown.
Interestingly, we observed that Salmonella SPI-1 effector SopB, which plays a crucial role in inducing uptake of bacteria inside non-phagocytic cells, inhibits the recruitment of various autophagic adaptor proteins such as Syntaxin 17 (STX17), microtubule-associated proteins 1A/1B light chain 3B (MAP1LC3B or LC3B), p62/SQSTM (sequestosome 1) onto the SCV membrane. Notably, we also found that SopB downregulates the transcript levels of these adaptor proteins and the overall autophagy flux inside host cells. We further demonstrate that SopB alters the PIP dynamics (PI3P and PI4P) of the SCV membrane. This activity of SopB helps the bacterium escape autophagy by inhibiting the fusion of SCVs with both lysosomes and autophagosomes.
Part II: To decipher the mechanism behind the reduction in lysosomal number in Salmonella-infected cells.
Previous work from our laboratory has shown that Salmonella infection causes a reduction in the overall lysosomal number inside the cell. This downregulation of lysosomal number facilitates survival of Salmonella inside host cells, as there are not enough lysosomes for the SCVs to fuse with and mediate clearance of the pathogen. While performing infection experiments with SopB mutants, we observed that the levels of LAMP1 and the number of lysosomes were significantly higher in SopB mutant infected cells than in STM WT. SopB has also been shown to activate Akt by phosphorylating Ser437 residue. In an independent study, it has been reported that Akt can further phosphorylate transcription factor EB (TFEB), which is a master regulator of the CLEAR (Coordinated Lysosomal Expression and Regulation) gene network. The phosphorylated TFEB is inactive and cannot translocate into the nucleus and hence cannot initiate the transcription of genes responsible for lysosomal biogenesis and autophagy. Therefore, we hypothesized that Salmonella SopB might also mediate overall downregulation through the Akt-TFEB axis in infected cells. We observed that SopB indeed reduced the number of acidic lysosomes inside the infected cells via the Akt-TFEB axis and thus facilitates the survival of Salmonella in host macrophages. These results were also validated in the mice model of Salmonella infection.
Part III: To elucidate the role of host Syntaxins in Salmonella pathogenesis and its virulence.
Intracellular membrane fusion is mediated by membrane-bridging complexes of soluble N-ethylmaleimide-sensitive factor attachment protein receptors (SNAREs). SNARE proteins are one of the key players in the endosomal trafficking pathways. Recent reports shed light on intracellular bacteria modulating host syntaxin to establish infection successfully. Salmonella actively modulates its vacuole to escape lysosomal fusion. One of the critical SNAREs in macrophages responsible for phagosome maturation is Syntaxin 3 and Syntaxin 4. Also, there is a report which suggests that SCV harbors Syntaxin 12. Therefore, in the third part of the study, based on the literature available, we have screened a few Syntaxins in SCV biogenesis and found that STX3 could play a role in SCV biogenesis. Upon knockdown of STX 3, we have observed that the bacterial proliferation is hindered and is restored upon the overexpression of STX3. We observed using live-cell imaging that during infection, SCV interacts with STX3 and thus might help in fusion and fission events of SCV with other vesicles to acquire membrane for facilitating the division of SCV. We also found this interaction abrogated when we infected with SPI-2 encoded T3SS apparatus mutant (STM ∆ssaV) but not with SPI-1 encoded T3SS (STM ∆invC). Together, these results indicated that the effector molecule secreted through SPI-2 encoded T3SS is involved in interaction with host STX3, which is essential to maintain Salmonella division along with SCV and maintenance of a single bacterium per vacuole.
Significant Findings and Conclusions:
This study concludes that Salmonella modulates host endo-lysosomal machinery to establish replicative niches inside host cells. Mainly, SopB plays a dual role in subverting the host cell xenophagy in macrophages; (1) by modulating the phosphatidylinositol phosphates dynamics on the SCV membrane to inhibit fusion of SCV with autophagosomes or lysosomes and (2) it downregulates the overall lysosomal biogenesis through Akt-TFEB axis inside infected host macrophages. This study also elucidates one of the critical interactions with host STX3, which helps the bacteria inside the host cell to divide along with SCV and is essential for maintaining a single bacterium per vacuole. Overall, we have deciphered that through the involvement of various effector molecules, Salmonella substantially modulates the host endo-lysosomal machinery to cause pathogenesis in the host.CSI
Insights into mRNA recruitment by eukaryotic initiation factor 4 during translation in Saccharomyces cerevisiae
No full textRibosomes are responsible for synthesizing proteins in the cells by using mRNA as the template. However, the ribosomes cannot perform this function without the help of other protein factors. Different factors are required at all the stages of translation – initiation factors during translation initiation, elongation factors during elongation, and termination factors are required for translation termination. Recycling factors then help in the dissociation of the ribosome into its large and small ribosomal subunits.
Of all the stages, initiation is a highly evolved, regulated, and complex process in eukaryotes compared to their prokaryotic counterparts. During initiation, mRNA is recruited to the small ribosomal subunit and is followed by recognition of the start codon. The Association of the large ribosomal subunit marks the end of initiation and results in the formation of the elongation competent ribosome. In prokaryotes, the base pairing of the Shine-Dalgarno sequence in mRNA with the anti-Shine-Dalgarno sequence present in the 16S rRNA leads to mRNA recruitment and places the start codon at the P-site. However, in the case of eukaryotes, there is a specialized group of factors called eIF4 that help in mRNA recruitment to the 40S. Studying the function of these eIF4 proteins in mRNA recruitment is the focus of this thesis.
eIF4B is one factor of the eIF4 group that is essential for the translation of mRNAs. Initial studies on this factor reported that it merely plays a stimulatory role during initiation, where it enhances the mRNA secondary structure unwinding activity of the RNA helicase factor eIF4A. However, there are also reports which show direct interaction of eIF4B with the 40S in yeast. To gain more insights into the roles played by eIF4B during initiation in yeast, we used cryo-EM to study the structure of the 40S-eIF4B complex. In the 4.8Å structure of 40S-eIF4B, we observed that eIF4B is located on the solvent-exposed side of the 40S and near the mRNA entry channel. Along with this, the mRNA channel latch is partially open when compared to the closed latch structure of the 40S. This suggests that eIF4B may be helping in mRNA recruitment by remodeling the 40S mRNA channel latch. When compared to the 43S PIC structure, we observe a steric clash of eIF4B with eIF3j and eIF3g-RRM. This clash may be responsible for the dissociation of eIF3j as well as the relocation of the eIF3b-g-i complex to the subunit interface of the 40S. With this study, we could decipher the location of eIF4B on the 40S, its interacting partners, and its contribution to the overall dynamics of translation initiation.
eIF4G & eIF4E are other factors of the eIF4 group where eIF4E recognizes the 5ʹ cap of the mRNA and eIF4G is the scaffolding protein responsible for keeping the entire eIF4 complex together. eIF4G also interacts with factors of the 43S PIC, thus helping in mRNA recruitment. This complex is also tightly regulated by the eIF4E and eIF4G binding proteins, keeping a check on mRNA recruitment. This regulation forms the basis for embryonic development, differentiation, aging, and the response of cells to stress conditions. Even after being important, limited structural knowledge of the eIF4G-eIF4E complex exists. This is mainly due to the presence of intrinsically disordered regions in eIF4G. In this work, we overexpressed & purified the S. cerevisiae eIF4G-eIF4E from E. coli and used cryo-EM to generate a 3D map of the complex. The complex appears to be L-shaped, however, the resolution is low owing to the conformational heterogeneity of eIF4G. Covalent cross-linking was tried to stabilize the complex, however, it resulted in aggregation of the sample during grid preparation. Overall, this study sheds light on the structural envelope of the eIF4G-eIF4E complex and highlights the technical challenges of studying the structure of this complex
Design, Control and Experimental Validation of a Robust Adaptive Feedback Linearization Controller for a Quadcopter Manipulator System
No full textA quadcopter manipulator system (QMS) is an aerial robot comprising a quadcopter with a
three degree of freedom manipulator mounted at the bottom of the vehicle. An aerial robot
can reach otherwise inaccessible locations, and the manipulator can execute a variety of tasks.
The presented thesis focuses on the design, control, and hardware implementation of a
quadcopter manipulator system, with a robust adaptive non-linear controller.
In the initial phase of the work, a thorough analysis of the workspace of the manipulator, which
is mounted at the bottom of the vehicle, is presented. The proposed robotic arm has an
extended workspace, allowing the end-effector to reach targets both above and below the
airframe. During tasks involving interaction with walls and other structures, the drone's thrust
can interact with the surroundings and cause counter moments on the system. A technique
based on drone and target positions is proposed to solve this problem.
A novel robust adaptive non-linear controller is designed and implemented in the second
phase. With uncertain time-varying parameters, the system has coupled non-linear dynamics. A
novel Augmented Adaptive Torque (AAT) control law is presented for the uncertain system,
which combines a model reference adaptive controller with a feedback linearization controller.
A strictly positive real-Lyapunov approach is used to create an adaptive law for estimating
unknown system parameters. Lyapunov theory is used to investigate the closed-loop system's
asymptotic stability. A bound on the parameter estimation error is derived utilizing the inputto-
state (ISS) stability concept.
The AAT control law is further combined with an estimate of the unknown bounded
disturbance to create the Robust Augmented Adaptive Torque (RAAT) control law, ensuring
robustness. The adaptive law is modified using a projection operator to ensure that the
estimates are bounded. To validate the theoretical conclusions and corroborate the
performance of the augmented adaptive torque control rule on the closed-loop system,
simulations in MATLAB and ROS/Gazebo are provided. A three DoF 3D printed robotic arm is
attached to an in-built quadcopter to create the QMS aerial robot, custom-built in the lab. To
assess the performance of the proposed controller, real-time experiments using QMS hardware
are carried out. The proposed method's efficiency is demonstrated by the aerial robot's
trajectory tracking and stability during real-time testing in field experiments
Correlation between Morphology, Microstructure and Corrosion Behaviour of Nickel-Phosphorous (Ni-P) Based Electrodeposited Coatings
No full textThe Ni-P alloy coatings are widely studied due to their superior mechanical and tribological properties. Ni-P coatings are also considered to be a viable alternative to the chromium (Cr) coatings which utilize environmentally hazardous and toxic carcinogenic electrolytic solutions. The current work focuses on strategies to enhance the corrosion resistance performance of electrodeposited Ni-P coatings primarily by incorporation of foreign additives (carbon nanotubes (CNTs) and graphene) and by engineering of the Ni-P micro-texture and phase fraction (crystalline and amorphous phases). Nickel-phosphorus (Ni-P) coatings were electrodeposited over mild steel substrate using DC power source in conventional two electrode electrochemical setup. As-deposited Ni-P coatings were subjected to phase, microstructural and morphological characterizations using x-ray diffraction, electron microscopy and electron backscatter diffraction techniques. The corrosion analysis was accomplished by using electrochemical impedance spectroscopy (EIS) and potentiodynamic polarisation techniques (Tafel plot) in 3.5 wt.% NaCl solution. Key observations are: (a) in the work on the incorporation of CNTs and Graphene in Ni-P coatings, it was observed that an optimum volume fraction of the additives yielded high corrosion resistance performance. This was essentially due to the smooth compact and defect free surface morphology, (b) in the work on the correlation between micro-texture and corrosion behaviour of Ni-P coatings as a function of phosphorous content, it was observed that the phosphorous concentration range where the nano-crystalline region (along with the minor amorphous phase fraction) was dominant a slight alteration in texture determines the corrosion rate. With increase in the amorphous region, the galvanic coupling between the anodic amorphous phase and cathodic crystalline phase determined the corrosion behaviour. A mixture of amorphous and crystalline phases with lower fraction of the amorphous phase enhanced the corrosion rate due to increased galvanic coupling. For higher addition of phosphorus, large fraction of amorphous phase evolved which significantly reduced the galvanic coupling leading to higher corrosion resistance behaviour, (c) in the work on the effect of deposition temperature (bath temperature of 15˚C, 20˚C, 25˚C, 35˚C) on the evolution of correlation between texture and corrosion behaviour of Ni-P coatings, it was observed that the coating deposited at 15°C and 25°C yielded the maximum and minimum corrosion rate respectively. Analysis of the coating texture revealed that the higher corrosion rate for the 15°C coating was due to low fraction of low energy low angle grain boundaries (LAGBs), higher strain within the grains, and (101) growth texture. Lower corrosion rate, on the other hand, for the 25°C coating was due to low energy (001) growth texture, low average strain within the grains, and high fraction of LAGBs, (d) in the work on the effect of deposition current density on the evolution of correlation between texture and corrosion behaviour of Ni-P coatings, it was observed that the Ni-P coating (deposited using 60 mA.cm-2) that exhibited the lowest corrosion rate was characterized by the presence of lower energy surface texture, lower grain size, narrow grain size distribution and a relatively higher fraction of low energy Σ3 coherent twin boundaries. A higher corrosion rate for coating deposited using 5 mA.cm-2 was due to higher energy surface texture and larger grain size distribution