No description available.

No description available.

No description available.

Self-organization of regular surface pattern under ion beam erosion was described in detail by Navez in 1962. Several years later in 1986 Bradley and Harper (BH) published the first self-consistent theory on this phenomenon based on the competition of surface roughening described by Sigmund's sputter theory and surface smoothing by Mullins-Herring diffusion. Many papers that followed BH theory introduced other processes responsible for the surface patterning e.g. viscous flow, redeposition, phase separation, preferential sputtering, etc. The present understanding is still not sufficient to specify the dominant driving forces responsible for self-organization. 3D atomistic simulations can improve the understanding by reproducing the pattern formation with the detailed microscopic description of the driving forces. 2D simulations published so far can contribute to this understanding only partially. A novel program package for 3D atomistic simulations called TRIDER (TRansport of Ions in matter with DEfect Relaxation), which unifies full collision cascade simulation with atomistic relaxation processes, has been developed. The collision cascades are provided by simulations based on the Binary Collision Approximation, and the relaxation processes are simulated with the 3D lattice kinetic Monte-Carlo method. This allows, without any phenomenological model, a full 3D atomistic description on experimental spatiotemporal scales. Recently discussed new mechanisms of surface patterning like ballistic mass drift or the dependence of the local morphology on sputtering yield are inherently included in our atomistic approach. The atomistic 3D simulations do not depend so much on experimental assumptions like reported 2D simulations or continuum theories. The 3D computer experiments can even be considered as 'cleanest' possible experiments for checking continuum theories. This work aims mainly at the methodology of a novel atomistic approach, showing that: (i) In general, sputtering is not the dominant driving force responsible for the ripple formation. Processes like bulk and surface defect kinetics dominate the surface morphology evolution. Only at grazing incidence the sputtering has been found to be a direct cause of the ripple formation. Bradley and Harper theory fails in explaining the ripple dynamics because it is based on the second-order-effect 'sputtering'. However, taking into account the new mechanisms, a 'Bradley-Harper equation' with redefined parameters can be derived, which describes pattern formation satisfactorily. (ii) Kinetics of (bulk) defects has been revealed as the dominating driving force of pattern formation. Constantly created defects within the collision cascade, are responsible for local surface topography fluctuation and cause surface mass currents. The mass currents smooth the surface at normal and close to normal ion incidence angles, while ripples appear first at incidence angles larger than 40°. The evolution of bimetallic interfaces under ion irradiation is another application of TRIDER described in this thesis. The collisional mixing is in competition with diffusion and phase separation. The irradiation with He ions is studied for two extreme cases of bimetals: (i) Irradiation of interfaces formed by immiscible elements, here Al and Pb. Ballistic interface mixing is accompanied by phase separation. Al and Pb nanoclusters show a self-ordering (banding) parallel to the interface. (ii) Irradiation of interfaces by intermetallics forming species, here Pt and Co. Well-ordered layers of phases of intermetallics appear in the sequence Pt/Pt3Co/PtCo/PtCo3/Co. The TRIDER program package has been proven to be an appropriate technique providing a complete picture of mixing mechanisms.

La construction du nid chez les fourmis génère des structures assumant de nombreuses fonctions, qui sont en grande partie dépendantes de l’architecture produite. L’omniprésence de ces fonctionnalités contraste avec la forte diversité intra et inter-spécifique de la morphologie du nid.<p>Ce travail a pour objectifs de déterminer d’une part la morphogenèse du nid durant son excavation par les fourmis (Chapitre 3), et d’autre part de quantifier l’impact de la taille du groupe (Chapitre 3) et de la qualité de l’environnement (qualité du matériau, gravité; Chapitres 4 & 5) sur cette séquence. Pour répondre à ces questions, nous avons utilisé un dispositif de creusement en deux dimensions (2D) afin de suivre la dynamique d’excavation et l’évolution de la morphologie du nid au cours du temps.<p>Nous avons tout d’abord mis en évidence (Chapitre 3) un changement brutal de la morphologie du nid au cours de sa croissance. Durant une première période d’excavation homogène, le nid est constitué d’une seule cavité de forme circulaire et au contour régulier. Par la suite, le pourtour de la cavité devient plus irrégulier, l’apparition de ‘‘bourgeons’’ lui donnant une apparence plus plissée. Enfin, la cavité centrale cesse de croître lorsque des ramifications se déploient depuis certains des ‘‘bourgeons’’. Nous avons qualifié de transitions morphologiques ces brusques changements de forme, dont la fréquence d’apparition augmente avec la taille de la population.<p>Notre analyse et le recours à la simulation semblent indiquer que ce phénomène soit basé non pas sur un changement de comportement des ouvrières excavatrices mais sur la densité d’activité au front de creusement. Lorsque celle-ci est importante, des phénomènes d’encombrement se manifestent et le nid est excavé de manière homogène. À l’inverse, lorsque l’encombrement diminue les ouvrières peuvent focaliser leur travail et excavent des galeries. Un modèle analytique (Chapitre 6) a permis d’étudier les conditions d’apparition de cette transition morphologique.<p>Nos autres résultats indiquent que le matériau de construction influence très fortement la morphogenèse du nid (Chapitre 4) :les nids excavés dans un milieu cohésif sont plus fréquemment ramifiés et la transition morphologique apparaît à des surfaces de nid plus petites que dans un milieu granulaire. Il semblerait qu’à nouveau, il n’y ait aucune variation comportementale impliquée, mais que l’environnement joue un rôle de ‘‘médiateur’’ des interactions entre les fourmis: le milieu change le temps d’extraction et donc la probabilité individuelle d’excavation, cette variation est exacerbée par les mécanismes d’amplification à l’œuvre, modifiant alors sensiblement la dynamique collective de creusement et la structure du nid.<p>Enfin, une série d’expériences (Chapitre 5) nous a permis de déterminer l’influence de la gravité sur la morphogenèse. Il apparait que la gravité sert de gabarit à la construction, puisque les nids excavés dans un dispositif à l’orientation verticale sont tous dirigés vers le bas. Aussi, si le nid prend directement une forme de long puit vertical sans jamais présenter de chambre, le mécanisme de transition morphologique reste visible, sous forme de bifurcation des galeries à leur extrémités. Ce phénomène de tip-splitting semble aussi apparaître sous la contrainte de la densité d’ouvrières excavant au fond des galeries.<p>Notre travail a donc mis en évidence un phénomène de transition morphologique capable de produire les modules de base de tout nid - chambres et galeries -, et qui semble découler non pas de modifications comportementales, mais de la seule interaction de l’amplification de l’activité et de l’encombrement au front d’excavation. Ce phénomène ubiquiste propose une explication à la grande variété de structures observées sur le terrain, en lien notamment avec la diversité et l’hétérogénéité des milieux, et la dynamique des colonies. Enfin, les similitudes de la transition morphologique avec les instabilités de croissance observées dans de nombreux systèmes biologiques et physico-chimique invitent à considérer ces derniers phénomènes à la lumière de nos résultats. / Doctorat en Sciences / info:eu-repo/semantics/nonPublished

We look at robotic systems made of separate discrete components that, by self-assembling, can organize into physical structures of growing size. We review 22 such systems, exhibiting components ranging from passive mechanical parts to mobile<p>robots. We present a taxonomy of the systems, and discuss their design and function. We then focus on a particular system, the swarm-bot. In swarm-bot, the components that assemble are self-propelled modules that are fully autonomous in power, perception, computation, and action. We examine the additional capabilities and functions self-assembly can offer an autonomous group of modules for the accomplishment of a concrete task: the transport of an object. The design of controllers is accomplished in simulation using<p>techniques from biologically-inspired computing. We show that self-assembly can offer adaptive value to groups that compete in an artificial evolution based on their fitness in task performance. Moreover, we investigate mechanisms that facilitate the design of self-assembling systems. The controllers are transferred to the physical swarm-bot system, and the capabilities of self-assembly and object transport are extensively evaluated in a range of different environments. Additionally, the controller for self-assembly is transferred and evaluated on a different robotic system, a super-mechano colony. Given the breadth and quality of the results obtained, we can say that the swarm-bot qualifies as the current state of the art in self-assembling robots. Our work supplies some initial evidence (in form of simulations and experiments with the swarm-bot) that self-assembly can offer robotic systems additional capabilities and functions useful for the accomplishment of concrete tasks.<p> / Doctorat en Sciences de l'ingénieur / info:eu-repo/semantics/nonPublished

Advances in unsupervised learning and deep neural networks have led to increased performance in a number of domains, and to the ability to draw strong comparisons between the biological method of self-organization conducted by the brain and computational mechanisms. This thesis aims to use real-world data to tackle two areas in the domain of computer vision which have biological equivalents: feature detection and motion tracking. The aforementioned advances have allowed efficient learning of feature representations directly from large sets of unlabeled data instead of using traditional handcrafted features. The first part of this thesis evaluates such representations by comparing regularization and preprocessing methods which incorporate local neighbouring information during training on a single-layer neural network. The networks are trained and tested on the Hollywood2 video dataset, as well as the static CIFAR-10, STL-10, COIL-100, and MNIST image datasets. The induction of topography or simple image blurring via Gaussian filters during training produces better discriminative features as evidenced by the consistent and notable increase in classification results that they produce. In the visual domain, invariant features are desirable such that objects can be classified despite transformations. It is found that most of the compared methods produce more invariant features, however, classification accuracy does not correlate to invariance. The second, and paramount, contribution of this thesis is a biologically-inspired model to explain the emergence of motion tracking behaviour in early development using unsupervised learning. The model’s self-organization is biased by an original concept called retinal constancy, which measures how similar visual contents are between successive frames. In the proposed two-layer deep network, when exposed to real-world video, the first layer learns to encode visual motion, and the second layer learns to relate that motion to gaze movements, which it perceives and creates through bi-directional nodes. This is unique because it uses general machine learning algorithms, and their inherent generative properties, to learn from real-world data. It also implements a biological theory and learns in a fully unsupervised manner. An analysis of its parameters and limitations is conducted, and its tracking performance is evaluated. Results show that this model is able to successfully follow targets in real-world video, despite being trained without supervision on real-world video.

This dissertation offers demonstrations of autonomous catalytic microtubes (microjet engines) with tunable diameters ranging from micro- to nanoscale and lengths from 50 μm to 1 mm. These results open the door to effective microengines and represent the entry in the Guinness Book of World Records for “the smallest man-made jet engine.” Several attractive methodologies of machine-based functions at the micro- and nanoscale are shown. For instance, catalytic Ti/Cr/Pt microjets, which are integrated on a planar substrate, can operate as “on chip” chemical micropumps by decomposition of hydrogen peroxide fuel into oxygen bubbles and water. When released from a substrate, microjets self-propel autonomously in solution. The incorporation of ferromagnetic layer (Fe) into the rolled-up geometry enables their remote control using external magnetic field. Such microjets were used to load, transport, deliver and assemble multiple cargo particles, including biological cells in bulk solutions and microfluidic channels. Furthermore, it is demonstrated that for microjets that are fixed to or self-propelled above a platinum patterned surface, the microengine power/speed can be controlled using a white lightsource. A change in intensity of the white light leads to a controllable switching “off” and “on” of the microengine power on demand. Light degrades a local concentration of the hydrogen peroxide fuel and surface tension and subsequently suppresses the generation of oxygen microbubbles. In the next step, the diameter of the microjets was rigorously reduced to 250 nm by using hybrid heteroepitaxial/catalytic InGaAs/GaAs/Cr/Pt nanotubes. Due to asymmetry of the rolled-up layers, these nanojets move in corkscrew-like motions and act as “self-propelled nanotools,” which were used in the next step to transport yeast cells and drill into fixed cancer Hela cells. Although it is well-known that hydrogen peroxide cannot be used to sustain viable cellular function, it is however conceivable that alternative fuels, such as glucose, might enable operation of such nanotools under biologically compatible conditions. As a first step to achieve this goal, demonstrations were made using metal-enzyme biocatalytic Ti/Au/SAM/Catalase microengines. Synthetic components with competing interactions are well-suited to study the emergence of their collective behavior, such as swarms of large numbers of individuals. Microengines’ self-organization in bistable swarms is shown at the air-liquid interface of the mixture of propylene carbonate and hydrogen peroxide. Microengines act as “water striders.” Buoyed by oxygen bubbles, they self-propel via the microbubble recoiling mechanism and, depending on the bubbles’ sizes, self-organize into swarms due to the meniscus climbing effect. These reversible swarms depend on the microengine power, which competes against attracting surface tension force. The demonstrated microjet engines show great promise for emerging applications, including biomedical, on-chip, environmental, and robotic micromachines. Furthermore, a key method discovered, entitled “rolled-up nanotechnology on polymers,” allowed for the fabrication of highly parallel arrays of microtubes with multiple functionalities and aimed for different purposes.

Aktive Filament-Systeme, wie zum Beispiel das Zellskelett, sind Beispiele einer interessanten Klasse neuartiger Materialien, die eine wichtige Rolle in der belebten Natur spielen. Viele wichtige Prozesse in lebenden Zellen wie zum Beispiel die Zellbewegung oder Zellteilung basieren auf dem Zellskelett. Das Zellskelett besteht aus Protein-Filamenten, molekularen Motoren und einer großen Zahl weiterer Proteine, die an die Filamente binden und diese zu einem Netz verbinden können. Die Filamente selber sind semifexible Polymere, typischerweise einige Mikrometer lang und bestehen aus einigen hundert bis tausend Untereinheiten, typischerweise Mono- oder Dimeren. Die Filamente sind strukturell polar, d.h. sie haben eine definierte Richtung, ähnlich einer Ratsche. Diese Polarität begründet unterschiedliche Polymerisierungs- und Depolymerisierungs-Eigenschaften der beiden Filamentenden und legt außerdem die Bewegungsrichtung molekularer Motoren fest. Die Polymerisation von Filamenten sowie Krafterzeugung und Bewegung molekularer Motoren sind aktive Prozesse, die kontinuierlich chemische Energie benötigen. Das Zellskelett ist somit ein aktives Gel, das sich fern vom thermodynamischen Gleichgewicht befindet. In dieser Arbeit präsentieren wir Beschreibungen solcher aktiven Filament-Systeme und wenden sie auf Strukturen an, die eine ähnliche Geometrie wie zellulare Strukturen haben. Beispiele solcher zellularer Strukturen sind Spannungsfasern, kontraktile Ringe oder mitotische Spindeln. Spannungsfasern sind für die Zellbewegung essentiell; sie können kontrahieren und so die Zelle vorwärts bewegen. Die mitotische Spindel trennt Kopien der Erbsubstanz DNS vor der eigentlichen Zellteilung. Der kontraktile Ring schließlich trennt die Zelle am Ende der Zellteilung. In unserer Theorie konzentrieren wir uns auf den Einfluß der Polymerisierung und Depolymerisierung von Filamenten auf die Dynamik dieser Strukturen. Wir zeigen, dass der kontinuierliche Umschlag (d.h. fortwährende Polymerisierung und Depolymerisierung) von Filamenten unabdingbar ist für die kontraktion eines Rings mit konstanter Geschwindigkeit, so wie in Experimenten mit Hefezellen beobachtet. Mit Hilfe einer mikroskopisch motivierten Beschreibung zeigen wir, wie &amp;quot;filament treadmilling&amp;quot;, also Filament Polymerisierung an einem Ende mit der gleichen Rate wie Depolymerisierung am anderen Ende, zur Spannung in Filament Bündeln und Ringen beitragen kann. Ein zentrales Ergebnis ist, dass die Depolymerisierung von Filamenten in Anwesenheit von filamentverbindenden Proteinen das Zusammenziehen dieser Bündel sogar in Abwesenheit molekulare Motoren herbeiführen kann. Ferner entwickeln wir eine generische Kontinuumsbeschreibung aktiver Filament-Systeme, die ausschließlich auf Symmetrien der Systeme beruht und von mikroskopischen Details unabhängig ist. Diese Theorie erlaubt uns eine komplementäre Sichtweise auf solche aktiven Filament-Systeme. Sie stellt ein wichtiges Werkzeug dar, um die physikalischen Mechanismen z.B. in Filamentbündeln aber auch bei der Bildung von Filamentringen im Zellkortex zu untersuchen. Schließlich entwickeln wir eine auf einem Kräftegleichgewicht basierende Beschreibung für bipolare Strukturen aktiver Filamente und wenden diese auf die mitotische Spindel an. Wir diskutieren Bedingungen für die Bildung und Stabilität von Spindeln. / Active filament systems such as the cell cytoskeleton represent an intriguing class of novel materials that play an important role in nature. The cytoskeleton for example provides the mechanical basis for many central processes in living cells, such as cell locomotion or cell division. It consists of protein filaments, molecular motors and a host of related proteins that can bind to and cross-link the filaments. The filaments themselves are semiflexible polymers that are typically several micrometers long and made of several hundreds to thousands of subunits. The filaments are structurally polar, i.e. they possess a directionality. This polarity causes the two distinct filament ends to exhibit different properties regarding polymerization and depolymerization and also defines the direction of movement of molecular motors. Filament polymerization as well as force generation and motion of molecular motors are active processes, that constantly use chemical energy. The cytoskeleton is thus an active gel, far from equilibrium. We present theories of such active filament systems and apply them to geometries reminiscent of structures in living cells such as stress fibers, contractile rings or mitotic spindles. Stress fibers are involved in cell locomotion and propel the cell forward, the mitotic spindle mechanically separates the duplicated sets of chromosomes prior to cell division and the contractile ring cleaves the cell during the final stages of cell division. In our theory, we focus in particular on the role of filament polymerization and depolymerization for the dynamics of these structures. Using a mean field description of active filament systems that is based on the microscopic processes of filaments and motors, we show how filament polymerization and depolymerization contribute to the tension in filament bundles and rings. We especially study filament treadmilling, an ubiquitous process in cells, in which one filament end grows at the same rate as the other one shrinks. A key result is that depolymerization of filaments in the presence of linking proteins can induce bundle contraction even in the absence of molecular motors. We extend this description and apply it to the mitotic spindle. Starting from force balance considerations we discuss conditions for spindle formation and stability. We find that motor binding to filament ends is essential for spindle formation. Furthermore we develop a generic continuum description that is based on symmetry considerations and independent of microscopic details. This theory allows us to present a complementary view on filament bundles, as well as to investigate physical mechanisms behind cell cortex dynamics and ring formation in the two dimensional geometry of a cylinder surface. Finally we present a phenomenological description for the dynamics of contractile rings that is based on the balance of forces generated by active processes in the ring with forces necessary to deform the cell. We find that filament turnover is essential for ring contraction with constant velocities such as observed in experiments with fission yeast.

Self-organization of regular surface pattern under ion beam erosion was described in detail by Navez in 1962. Several years later in 1986 Bradley and Harper (BH) published the first self-consistent theory on this phenomenon based on the competition of surface roughening described by Sigmund's sputter theory and surface smoothing by Mullins-Herring diffusion. Many papers that followed BH theory introduced other processes responsible for the surface patterning e.g. viscous flow, redeposition, phase separation, preferential sputtering, etc. The present understanding is still not sufficient to specify the dominant driving forces responsible for self-organization. 3D atomistic simulations can improve the understanding by reproducing the pattern formation with the detailed microscopic description of the driving forces. 2D simulations published so far can contribute to this understanding only partially. A novel program package for 3D atomistic simulations called TRIDER (TRansport of Ions in matter with DEfect Relaxation), which unifies full collision cascade simulation with atomistic relaxation processes, has been developed. The collision cascades are provided by simulations based on the Binary Collision Approximation, and the relaxation processes are simulated with the 3D lattice kinetic Monte-Carlo method. This allows, without any phenomenological model, a full 3D atomistic description on experimental spatiotemporal scales. Recently discussed new mechanisms of surface patterning like ballistic mass drift or the dependence of the local morphology on sputtering yield are inherently included in our atomistic approach. The atomistic 3D simulations do not depend so much on experimental assumptions like reported 2D simulations or continuum theories. The 3D computer experiments can even be considered as 'cleanest' possible experiments for checking continuum theories. This work aims mainly at the methodology of a novel atomistic approach, showing that: (i) In general, sputtering is not the dominant driving force responsible for the ripple formation. Processes like bulk and surface defect kinetics dominate the surface morphology evolution. Only at grazing incidence the sputtering has been found to be a direct cause of the ripple formation. Bradley and Harper theory fails in explaining the ripple dynamics because it is based on the second-order-effect 'sputtering'. However, taking into account the new mechanisms, a 'Bradley-Harper equation' with redefined parameters can be derived, which describes pattern formation satisfactorily. (ii) Kinetics of (bulk) defects has been revealed as the dominating driving force of pattern formation. Constantly created defects within the collision cascade, are responsible for local surface topography fluctuation and cause surface mass currents. The mass currents smooth the surface at normal and close to normal ion incidence angles, while ripples appear first at incidence angles larger than 40°. The evolution of bimetallic interfaces under ion irradiation is another application of TRIDER described in this thesis. The collisional mixing is in competition with diffusion and phase separation. The irradiation with He ions is studied for two extreme cases of bimetals: (i) Irradiation of interfaces formed by immiscible elements, here Al and Pb. Ballistic interface mixing is accompanied by phase separation. Al and Pb nanoclusters show a self-ordering (banding) parallel to the interface. (ii) Irradiation of interfaces by intermetallics forming species, here Pt and Co. Well-ordered layers of phases of intermetallics appear in the sequence Pt/Pt3Co/PtCo/PtCo3/Co. The TRIDER program package has been proven to be an appropriate technique providing a complete picture of mixing mechanisms.