12023 research outputs found
Sort by
Locomotory Control Algorithms and Their Neuronal Implementation in Drosophila melanogaster
Get PDFScientists and engineers alike have long looked to animals in their pursuit of understanding the natural world and how best to interact with it. While researchers have looked across diverse classes, insects have been extensively studied for their rich diversity of life histories and abilities to perform at spatial and temporal scales difficult for engineered systems. Within insects, the fruit fly, Drosophila melanogaster, is a particularly well-studied organism because of its experimental tractability and status as a genetic model organism, providing both detailed descriptions of a broad suite of behaviors and access to and control over specific sets of tissue. In this work, we make use of these tools to study two behaviors in Drosophila, local search, the behavior in which walking flies will search the area around a food site in search of other food sources nearby, and the optomotor response, wherein they will stabilize in response to visual motion during flight. In these studies, we will use modern techniques from both biology and engineering, to exhaustively characterize and describe the observed behaviors and attempt to untangle the underlying algorithms and their neuronal implementation.
First, we explore the algorithmic structure of local search in fruit flies. When flies encounter a piece of food, they will often perform walking searches nearby; as food tends to be patchy in natural settings, searches may allow flies to locate other food sites in the area. We induce local search using optogenetic stimulation of sugar-sensing neurons and constrain the flies to a dark, annular arena. These experimental details result in a simplified behavior, as the fly has access to a limited sensory environment, so that the search can be interpreted as an example of idiothetic path integration, and the search itself is one-dimensionalized and therefore more easily analyzed. Our experiments, in tandem with complementary modeling using a state transition diagram formalization of the behaviors, generate two principle findings. First, flies can integrate their location in two dimensions--after the optogenetic activation is disabled and the flies can no longer receive the food stimulus, they will continue to search over the former food site even after completing a full revolution of the annular arena. Second, when multiple food sites are present, they search over a center of the food sites, rather than over one distinct food site. These results both provide insights into the algorithmic structure of local search and an experimental and descriptive paradigm for further inquiries into the behavior.
Second, we investigate the role of a population of neurons, the DNg02s, in the optomotor response. In response to visual patterns of wide-field motion, such that the entire world is moving in the fly-centric reference frame, the animal will attempt to steer to cancel the visual motion, as the most parsimonious explanation of the motion is that the fly itself is moving in the global reference frame. We demonstrate that the DNg02 neurons are a required component in the neural circuitry underlying the optomotor response, but that they are insufficient to induce steering behaviors. We conclude with a set of models that fully recapitulate the collected dataset. With current techniques, distinguishing between the two possible models of the downstream connectivity from the DNg02s to the motor neurons associated with wing motor output is not possible. However, as new datasets become available, particularly complete connectomes of the Drosophila nervous system, the neuronal pathways from the DNg02s to the motor systems may be elucidated.</p
Contraction Theory for Robust Learning-Based Control: Toward Aerospace and Robotic Autonomy
Get PDFMachine learning and AI have been used for achieving autonomy in various aerospace and robotic systems. In next-generation research tasks, which could involve highly nonlinear, complicated, and large-scale decision-making problems in safety-critical situations, however, the existing performance guarantees of black-box AI approaches may not be sufficiently powerful. This thesis gives a mathematical overview of contraction theory, with some practical examples drawn from joint projects with NASA JPL, for enjoying formal guarantees of nonlinear control theory even with the use of machine learning-based and data-driven methods. This is not to argue that these methods are always better than conventional approaches, but to provide formal tools to investigate their performance for further discussion, so we can design and operate truly autonomous aerospace and robotic systems safely, robustly, adaptively, and intelligently in real-time.
Contraction theory is an analytical tool to study differential dynamics of a non-autonomous (i.e., time-varying) nonlinear system under a contraction metric defined with a uniformly positive definite matrix, the existence of which results in a necessary and sufficient characterization of incremental exponential stability of multiple solution trajectories with respect to each other. Its nonlinear stability analysis boils down to finding a suitable contraction metric that satisfies a stability condition expressed as a linear matrix inequality, resulting in many parallels drawn between linear systems theory and contraction theory for nonlinear systems. This yields much-needed safety and stability guarantees for neural network-based control and estimation schemes, without resorting to a more involved method of using uniform asymptotic stability for input-to-state stability. Such distinctive features permit the systematic construction of a contraction metric via convex optimization, thereby obtaining an explicit exponential bound on the distance between a time-varying target trajectory and solution trajectories perturbed externally due to disturbances and learning errors. The first two parts of this thesis are about a theoretical overview of contraction theory and its advantages, with an emphasis on deriving formal robustness and stability guarantees for deep learning-based 1) feedback control, 2) state estimation, 3) motion planning, 4) multi-agent collision avoidance and robust tracking augmentation, 5) adaptive control, 6) neural net-based system identification and control, for nonlinear systems perturbed externally by deterministic and stochastic disturbances. In particular, we provide a detailed review of techniques for finding contraction metrics and associated control and estimation laws using deep neural networks.
In the third part of the thesis, we present several numerical simulations and empirical validation of our proposed approaches to assess the impact of our findings on realizing aerospace and robotic autonomy. We mainly focus on the two joint projects with NASA JPL: 1) Science-Infused Spacecraft Autonomy for Interstellar Object Exploration and 2) Constellation Autonomous Space Technology Demonstration of Orbital Reconfiguration (CASTOR), where we also perform hardware demonstrations of our methods using our thruster-based spacecraft simulators (M-STAR) and in high-conflict, distributed, intelligent UAV swarm reconfiguration with up to 20 UAVs (crazyflies).</p
Expanding the Toolbox for Thermal Control of E. coli: Cold-Activated Transcription with Applications in Temperature Self-Regulation
Get PDFTemperature can be used to control engineered E. coli — for example, the living component of an engineered living material (ELM) - through the use of thermolabile transcription factors. Sharp induction of gene expression with heat has been established using these bacteria- and phage-derived proteins. Here, we expand the toolbox for thermal control of E. coli through both direct cold-induced gene expression and through the construction of genetic circuits to invert heat-induced gene expression.
We accomplish direct induction at low temperatures through the use of temperature-sensitive mutants of Lambda repressor as transcriptional activators. In addition, we show that a temperature-sensitive mutant of Lambda repressor can serve as an activator and a repressor of different genes simultaneously in one genetic circuit, leading to opposite thermal responses and serving as a temperature switch.
Next, we demonstrate inversion of a temperature-sensitive repressor using a temperature insensitive repressor. We apply this multicomponent switch to engineer a temperature self-regulation circuit for E. coli-based ELMs. Seasonal variation in ambient temperature presents a challenge in deploying ELMs outside of a laboratory environment, because E. coli growth rate is impaired both below and above 37°C. Our construct enables E. coli to produce a light-absorptive pigment in response to environmental temperature below 36°C with the goal of allowing the cells to absorb sunlight and locally warm to their optimal growth temperature. We demonstrate the efficacy of our pigment temperature switch in a model flat ELM growing at 32°C and 42°C in a home-built illuminated growth chamber. Below 36°C, our engineered E. coli increase in pigmentation, causing an increase in sample temperature and growth rate above non-pigmented bacteria. On the other hand, above 36°C, they decrease in pigmentation, protecting their growth compared to bacteria with temperature- independent high pigmentation. Integrating our temperature homeostasis circuit into an ELM has the potential to improve ELM performance by optimizing growth and protein production in the face of seasonal temperature changes.</p
Acoustic Biomolecules for Diagnostic Ultrasound Imaging
Get PDFNanotechnology has enabled significant breakthroughs in the early detection and treatment of disease, but many of these advances rely on expensive and less-accessible imaging modalities. Ultrasound, on the other hand, is a noninvasive imaging modality that stands out for its universal availability, cost-effectiveness, and safety. However, harnessing the benefits of nanomaterials for ultrasound has been challenging due to the size and stability constraints of typical ultrasound contrast agents. Recently, an innovative solution has emerged in the form of gas vesicles (GVs), a class of air-filled protein nanostructures found in certain aquatic microbes. These promising next-generation ultrasound contrast agents offer a crucial bridge between nanotechnology and ultrasonography.
In this thesis, we investigate the in vivo behavior of GVs, explore their potential applications as nanodiagnostic agents, and consider key factors for their future clinical deployment. In Chapter 2, we examine the interactions of GVs with blood components, focusing on imaging performance and immunogenicity. In Chapter 3, we show that intravenously injected GVs are cleared by liver-resident macrophages and subsequently undergo lysosomal degradation. We leverage this finding to develop an ultrasound-based method for visualizing cellular degradative processes and demonstrate its potential as a liver disease diagnostic. In Chapter 4, we introduce bicone GVs, the smallest known ultrasound contrast agent. We show that these sub-80 nm particles can penetrate tumors, deliver potent ultrasound-induced mechanical effects, and are readily engineered for molecular targeting, extended circulation time, and payload conjugation.
Together, these findings highlight the tremendous potential of GVs as injectable nanomaterials for ultrasound imaging, laying the foundation for future studies to further refine the design and application of these agents.</p
Laser Spectroscopy of Hydrocarbons for Applications in Atmospheric and Space Science
Get PDFThis thesis describes applications of spectroscopy and mass spectrometry towards applications of in situ sensing, chemical kinetics, and photodissociation processes of hydrocarbon species. Both mass spectrometry and cavity ring-down spectroscopy are used in this work. Irradiation of protonated coronene was used to study photodissociation process of polycyclic aromatic hydrocarbons (PAHs), with the power, irradiation time, and fragmentation studied to elucidate a photofragmentation mechanism. Biomolecules were ionized via Direct Analysis in Real Time (DART) to explore the feasibility and sensitivity of the ionization technique for different chemical species for in situ measurement in simulated extraterrestrial conditions. Photofragmentation and DART ionization were combined to quantify mixtures of isobaric PAHs, providing a tunable and complimentary technique for in situ analysis of mixtures. Finally, frequency-stabilized cavity ring-down spectroscopy was used to analyze precise ¹³CH₄, CH₃D, and CH₂D₂ to CH₄ isotologue ratios using optically-switched dual-wavelengths, allowing for sensitive measurement of the kinetic isotope effect of methane oxidation with O(¹D).</p
Stochastic Foundations for Single-Cell RNA Sequencing
Get PDFSingle-cell RNA sequencing, which quantifies cell transcriptomes, has seen widespread adoption, accompanied by proliferation of analysis methods. However, there has been relatively little systematic investigation of its best practices and their underlying assumptions, leading to challenges and discrepancies in interpretation. I present a set of generic, principled strategies for modeling the biological and technical components of sequencing experiments and use case studies to motivate their application to sequencing data.</p
Advancements in Hemodynamic Measurement: Arterial Resonance, Ultrasound, and Machine Learning
Get PDFThis thesis covers two separate projects which both use ultrasound to measure a form of blood pressure in very different ways. The first project focuses on the noninvasive measurement of continuous arterial blood pressure via the previously unstudied phenomenon of arterial resonance. While prior research efforts have attempted many methods of noninvasive blood pressure measurement, none has been able to generate continuous, calibration-free measurements based on a first-principles physical model. This work describes the derivation of this resonance-based model, its in vitro validation, and its in vivo testing on 60 subjects. This testing resulted in robust resonance detection and accurate calculation of BP in the large majority of evaluated subjects, representing very promising performance for the first test of a new biomedical technology. The second study changes focus to the measurement of blood pressure in the right atrium of the heart, an important clinical indicator in heart disease patients. Rather than developing a new physical approach, this project used machine learning to model the existing assessments made by cardiologists. Comparison to gold standard invasive catheter measurements showed that model predictions were statistically indistinguishable from cardiologist measurements. Both of these projects represent significant advances in expanding precise blood pressure measurements beyond critical care units and expanding access to a much broader population.</p
Ultraviolet Scalar Unification and Gravitational Radiation Reaction from Quantum Field Theory
Get PDFSince its inception, the development of quantum field theory has been driven by a desire to describe nature at the highest energy scales, where both relativistic and quantum mechanical aspects of matter and radiation are manifest. The theory has been wildly successful in this respect, giving rise to the standard model of particle physics, as well as quantum cosmologies of the early universe. The applications of quantum field theory are not, however, restricted to high-energy physics. The theory is just as spectacular in the infrared as it is in the ultraviolet, and it serves as a mathematical nexus for physical processes spanning the energy spectrum.
We will investigate two such connections in this work. In the first part, we will relate the analytic structure of the scattering amplitudes of scalar quantum field theories in the far infrared to unifying symmetries of their actions in the ultraviolet. In the sequel, we will study a system of two massive scalar fields coupled to the spacetime metric. Identifying the classical limit with the gravitational infrared, we will sift the classical dynamics of binary gravitational inspiral from the scattering amplitudes of canonical quantum gravity.
In Chapter 1, we argue that symmetry and unification can emerge as byproducts of certain physical constraints on dynamical scattering. To accomplish this we parameterize a general Lorentz invariant, four-dimensional theory of massless and massive scalar fields coupled via arbitrary local interactions. Assuming perturbative unitarity and an Adler zero condition, we prove that any finite spectrum of massless and massive modes will necessarily unify at high energies into multiplets of a linearized symmetry. Certain generators of the symmetry algebra can be derived explicitly in terms of the spectrum and three-particle interactions. Furthermore, our assumptions imply that the coset space is symmetric.
In Chapter 2, we introduce the gravitational inspiral problem.In Chapter 3, we review the Hamiltonian formulation of binary dynamics and outline the extraction of the effective Hamiltonian from scattering amplitudes. We then augment the equations of motion with a gravitational radiation reaction force, thus incorporating dissipation. In Chapter 4, we extend the setup of Kosower, Maybee, and O'Connell (KMOC), which expresses classical observables in terms of scattering amplitudes, to study the evolution of angular momentum during two-body scattering in gravity and electromagnetism. From the associated scattering amplitudes, we explicitly compute the total radiated angular momentum through the third Post-Minkowskian order (3PM) in general relativity and the third Post-Coulombic order (3PC) in electromagnetism.
In Chapter 5, starting with the classical expressions for radiated energy and angular momentum, we compute the corresponding instantaneous radiative fluxes in isotropic gauge. We then use these fluxes to derive the relative radiation reaction force associated to a gravitational binary system at third post-Minkowskian order by imposing flux balance. Together with the conservative Hamiltonian, this force provides a complete equation of motion for the relative degree of freedom of a (bound or unbound) gravitational binary at 3PM. Finally, in Chapter 6 we compare our results with the post-Newtonian literature, finding agreement to G³, 3PN (relative) order.</p
Listening to the Internal Representation of Actions Within the Posterior Parietal Cortex
Get PDFMore than 5.4 million people in the United States live with chronic paralysis and roughly 20 million people worldwide live with spinal cord injuries. Brain-machine interfaces (BMIs) can be transformative for these people, enabling them to control computers, robots, and more with only thought. State-of-the-art BMIs have already made this future a reality in limited clinical trials. However, these state-of-the-art BMIs have shortcomings that limit user adoption; high-performance BMIs currently require highly invasive electrodes above or in the brain; device degradation limits longevity to about 5 years; and their field of view is small, restricting the number, and type, of applications possible. This illustrates the need for a new generation of BMIs with a brain recording modality that is longer lasting, less invasive, and scalable to sense activity from large regions of the brain.
Functional ultrasound imaging (fUSI) is a recently developed technique that meets these criteria. fUSI measures cerebral hemodynamics with exceptional spatiotemporal resolution (<100 µm; ~100 ms) and a large field of view (several cm)—specifications ideally suited to recording detailed activity of entire cortical regions in parallel. In a series of novel results, we work towards developing the first high-performance ultrasonic BMI for human use. We first demonstrate that posterior parietal cortex (PPC), an area important for sensorimotor transformation, contains mesoscopic populations tuned to the intended movement direction. Using offline recorded data from several rhesus macaque monkeys, we can decode intended movement direction, task state, and expected action reward magnitude on a single trial basis. Having demonstrated that we could decode a variety of motor and cognitive variables using offline data, we developed a real-time, closed-loop ultrasonic BMI capable of decoding up to eight directions of intended movement with high accuracy. Finally, we began to translate these results into human applications and demonstrate the ability to measure changes in cerebral hemodynamics with high sensitivity through an acoustically transparent skull replacement in human subjects.
Taken together, our work is a novel characterization of how functional ultrasound neuroimaging may enable a new generation of BMIs. Additionally, this work reinforces the validity of fUSI as a robust and accessible neuroimaging technique for future neuroscience questions about mesoscopic populations and their interrelationships throughout the brain.</p
Wave Propagation in Periodic Acoustic Metamaterials: from 1D to 3D
Get PDFWave propagation in periodic structures has been studied for centuries; for example, Newton derived the velocity of sound based on a linear lattice. Recently, advanced manufacturing techniques have led to the fabrication of geometrically complex architected materials with acoustic properties unattainable by their constituent materials. Such rationally designed structures are often called acoustic metamaterials and they can be engineered to transmit, block, amplify, or redirect acoustic waves. Subwave-length building blocks, typically periodic (but not necessarily so), can be assembled into effectively continuous materials to manipulate dispersive properties of vibrational waves in ways that differ substantially in conventional media. This thesis investigates rationally designed acoustic metamaterials, ranging from 1D to 3D, and how acoustic wave propagation can be controlled by these artificially structured composite materials for ultrasound-related biomedical applications.
I first explore 1D wave propagation in acoustic metamaterials to study the basic mechanics and relevant analysis skills. Bio-inspired helical mechanical metamaterials are designed and their normal modes are investigated. I demonstrate the ability to vary the acoustic properties of the helical metamaterials by perturbing the geometrical structure and mass distribution. By locally adding eccentric and denser elements in the unit cells, I change the moment of inertia of the system and introduce centro-asymmetry. This allows me to control the degree of mode coupling and the width of subwavelength band gaps in the dispersion relation, which are the product of enhanced local resonance hybridization.
Then I study 2D wave propagation in microlattice acoustic metamaterials for ultra- sound manipulation. When coupled with pressure waves in the surrounding fluid, the dynamic behavior of microlattices in the long wavelength limit can be explained in the context of Biot’s theory of poroelasticity. I exploit elastoacoustic wave propagation within 3D-printed polymeric microlattices to design a gradient refractive index lens for underwater wave focusing. A modified Luneburg lens index profile adapted for ultrasonic wave lensing is demonstrated via the finite element method and underwater testing, showcasing a computationally efficient poroelasticity-based design approach that enables accelerated design of acoustic wave manipulation devices.
Lastly, I show that tailorable 3D wave propagation can be achieved based on the findings from the previous chapters. Functional ultrasound imaging enables sensitive, high-resolution imaging of neural activity in freely behaving animals and human patients. However, the skull acts as an aberrant and absorbing layer for sound waves, leading to most functional ultrasound experiments being conducted after skull removal. A microscale 2-photon polymerization technique is adopted to fabricate a conformal acoustic window with a high stiffness-to-density ratio and sonotransparency. Long-term biocompatibility and lasting signal sensitivity are demonstrated over a long period of time (> 4 months) by conducting ultrasound imaging in mouse models implanted with the metamaterial skull prosthesis.</p