Processos aleatórios não-markovianos: perfis de memória

One of the mechanisms responsible for the anomalous diffusion is the existence of long-range temporal correlations, for example, Fractional Brownian Motion and walk models according to Elephant memory and Alzheimer profiles, whereas in the latter two cases the walker can always "remember" of his first steps. The question to be elucidated, and the was the main motivation of our work, is if memory of the historic initial is condition for observation anomalous diffusion (in this case, superdiffusion). We give a conclusive answer, by studying a non-Markovian model in which the walkers memory of the past, at time t, is given by a Gaussian centered at time t=2 and standard deviation t which grows linearly as the walker ages. For large widths of we find that the model behaves similarly to the Elephant model; In the opposite limit (! 0), although the walker forget the early days, we observed similar results to the Alzheimer walk model, in particular the presence of amnestically induced persistence, characterized by certain log-periodic oscillations. We conclude that the memory of earlier times is not a necessary condition for the generating of superdiffusion nor the amnestically induced persistence and can appear even in profiles of memory that forgets the initial steps, like the Gausssian memory profile investigated here.

Odometria visual baseada em técnicas de structure from motion

Visual Odometry is the process that estimates camera position and orientation based solely on images and in features (projections of visual landmarks present in the scene) extraced from them. With the increasing advance of Computer Vision algorithms and computer processing power, the subarea known as Structure from Motion (SFM) started to supply mathematical tools composing localization systems for robotics and Augmented Reality applications, in contrast with its initial purpose of being used in inherently offline solutions aiming 3D reconstruction and image based modelling. In that way, this work proposes a pipeline to obtain relative position featuring a previously calibrated camera as positional sensor and based entirely on models and algorithms from SFM. Techniques usually applied in camera localization systems such as Kalman filters and particle filters are not used, making unnecessary additional information like probabilistic models for camera state transition. Experiments assessing both 3D reconstruction quality and camera position estimated by the system were performed, in which image sequences captured in reallistic scenarios were processed and compared to localization data gathered from a mobile robotic platform

Um estudo na teoria do movimento brownianno com viscosidade variável

In this work we investigate the stochastic behavior of a large class of systems with variable damping which are described by a time-dependent Lagrangian. Our stochastic approach is based on the Langevin treatment describing the motion of a classical Brownian particle of mass m. Two situations of physical interest are considered. In the first one, we discuss in detail an application of the standard Langevin treatment (white noise) for the variable damping system. In the second one, a more general viewpoint is adopted by assuming a given expression to the so-called collored noise. For both cases, the basic diffententiaql equations are analytically solved and al the quantities physically relevant are explicitly determined. The results depend on an arbitrary q parameter measuring how the behavior of the system departs from the standard brownian particle with constant viscosity. Several types of sthocastic behavior (superdiffusive and subdiffusive) are obteinded when the free pamameter varies continuosly. However, all the results of the conventional Langevin approach with constant damping are recovered in the limit q = 1