A study on tackling visual odometry by a transformer architecture

This dissertation describes a deepening study about Visual Odometry problem tackled with transformer architectures. The existing VO algorithms are based on heavily hand-crafted features and are not able to generalize well to new environments. To train them, we need carefully fine-tune the hyper-parameters and the network architecture. We propose to tackle the VO problem with transformer because it is a general-purpose architecture and because it was designed to transformer sequences of data from a domain to another one, which is the case of the VO problem. Our first goal is to create synthetic dataset using BlenderProc2 framework to mitigate the problem of the dataset scarcity. The second goal is to tackle the VO problem by using different versions of the transformer architecture, which will be pre-trained on the synthetic dataset and fine-tuned on the real dataset, KITTI dataset. Our approach is defined as follows: we use a feature-extractor to extract features embeddings from a sequence of images, then we feed this sequence of embeddings to the transformer architecture, finally, an MLP is used to predict the sequence of camera poses.

SmartForge: An event-driven architecture for simulator-based process optimization

In the metal industry, and more specifically in the forging one, scrap material is a crucial issue and reducing it would be an important goal to reach. Not only would this help the companies to be more environmentally friendly and more sustainable, but it also would reduce the use of energy and lower costs. At the same time, the techniques for Industry 4.0 and the advancements in Artificial Intelligence (AI), especially in the field of Deep Reinforcement Learning (DRL), may have an important role in helping to achieve this objective. This document presents the thesis work, a contribution to the SmartForge project, that was performed during a semester abroad at Karlstad University (Sweden). This project aims at solving the aforementioned problem with a business case of the company Bharat Forge Kilsta, located in Karlskoga (Sweden). The thesis work includes the design and later development of an event-driven architecture with microservices, to support the processing of data coming from sensors set up in the company's industrial plant, and eventually the implementation of an algorithm with DRL techniques to control the electrical power to use in it.

Sequencing cutting patterns with colored interval graphs

The problem addressed here originates in the industry of flat glass cutting and wood panel sawing, where smaller items are cut from larger items accordingly to predefined cutting patterns. In this type of industry the smaller pieces that are cut from the patterns are piled around the machine in stacks according to the size of the pieces, which are moved to the warehouse only when all items of the same size have been cut. If the cutting machine can process only one pattern at a time, and the workspace is limited, it is desirable to set the sequence in which the cutting patterns are processed in a way to minimize the maximum number of open stacks around the machine. This problem is known in literature as the minimization of open stacks (MOSP). To find the best sequence of the cutting patterns, we propose an integer programming model, based on interval graphs, that searches for an appropriate edge completion of the given graph of the problem, while defining a suitable coloring of its vertices.

Flexible multivariate GARCH modeling with an application to international stock markets

The goal of this paper is to estimate time-varying covariance matrices.Since the covariance matrix of financial returns is known to changethrough time and is an essential ingredient in risk measurement, portfolioselection, and tests of asset pricing models, this is a very importantproblem in practice. Our model of choice is the Diagonal-Vech version ofthe Multivariate GARCH(1,1) model. The problem is that the estimation ofthe general Diagonal-Vech model model is numerically infeasible indimensions higher than 5. The common approach is to estimate more restrictive models which are tractable but may not conform to the data. Our contributionis to propose an alternative estimation method that is numerically feasible,produces positive semi-definite conditional covariance matrices, and doesnot impose unrealistic a priori restrictions. We provide an empiricalapplication in the context of international stock markets, comparing thenew estimator to a number of existing ones.

Joint rolling-horizon scheduling of materials processing and lot-sizing with sequence-dependent setups

A lot sizing and scheduling problem from a foundry is considered in which key materials are produced and then transformed into many products on a single machine. A mixed integer programming (MIP) model is developed, taking into account sequence-dependent setup costs and times, and then adapted for rolling horizon use. A relax-and-fix (RF) solution heuristic is proposed and computationally tested against a high-performance MIP solver. Three variants of local search are also developed to improve the RF method and tested. Finally the solutions are compared with those currently practiced at the foundry.

The application of preprocessing and cutting plane techniques for a class of production planning problems

This paper investigates properties of integer programming models for a class of production planning problems. The models are developed within a decision support system to advise a sales team of the products on which to focus their efforts in gaining new orders in the short term. The products generally require processing on several manufacturing cells and involve precedence relationships. The cells are already (partially) committed with products for stock and to satisfy existing orders and therefore only the residual capacities of each cell in each time period of the planning horizon are considered. The determination of production recommendations to the sales team that make use of residual capacities is a nontrivial optimization problem. Solving such models is computationally demanding and techniques for speeding up solution times are highly desirable. An integer programming model is developed and various preprocessing techniques are investigated and evaluated. In addition, a number of cutting plane approaches have been applied. The performance of these approaches which are both general and application specific is examined.

Generating unconstrained two-dimensional non-guillotine cutting patterns by a Recursive Partitioning Algorithm

In this study, a dynamic programming approach to deal with the unconstrained two-dimensional non-guillotine cutting problem is presented. The method extends the recently introduced recursive partitioning approach for the manufacturer's pallet loading problem. The approach involves two phases and uses bounds based on unconstrained two-staged and non-staged guillotine cutting. The method is able to find the optimal cutting pattern of a large number of pro blem instances of moderate sizes known in the literature and a counterexample for which the approach fails to find known optimal solutions was not found. For the instances that the required computer runtime is excessive, the approach is combined with simple heuristics to reduce its running time. Detailed numerical experiments show the reliability of the method. Journal of the Operational Research Society (2012) 63, 183-200. doi: 10.1057/jors.2011.6 Published online 17 August 2011

Implications of the reduction of cutting fluid in drilling AISI P20 steel with carbide tools

The machining of hardened steel is becoming increasingly important in manufacturing processes. Machined parts made with hardened steel are often subjected to high service demands, which require great resistance and quality. The machining of this material submits the tools to high mechanical and thermal loads, which increases the tool wear and affects the surface integrity of the part. In that context, this work presents a study of drilling of AISI P20 steel with carbide tools, analyzing the effects on the process caused by the reduction of cutting fluid supply and its relation with the tool wear and the surface integrity of the piece. The major problem observed in the tests was a difficulty for chips to flow through the drill flute, compromising their expulsion from the hole. After a careful analysis, a different machining strategy was adopted to solve the problem

On the Filtering Problem for Continuous-Time Markov Jump Linear Systems with no Observation of the Markov Chain

We consider in this paper the optimal stationary dynamic linear filtering problem for continuous-time linear systems subject to Markovian jumps in the parameters (LSMJP) and additive noise (Wiener process). It is assumed that only an output of the system is available and therefore the values of the jump parameter are not accessible. It is a well known fact that in this setting the optimal nonlinear filter is infinite dimensional, which makes the linear filtering a natural numerically, treatable choice. The goal is to design a dynamic linear filter such that the closed loop system is mean square stable and minimizes the stationary expected value of the mean square estimation error. It is shown that an explicit analytical solution to this optimal filtering problem is obtained from the stationary solution associated to a certain Riccati equation. It is also shown that the problem can be formulated using a linear matrix inequalities (LMI) approach, which can be extended to consider convex polytopic uncertainties on the parameters of the possible modes of operation of the system and on the transition rate matrix of the Markov process. As far as the authors are aware of this is the first time that this stationary filtering problem (exact and robust versions) for LSMJP with no knowledge of the Markov jump parameters is considered in the literature. Finally, we illustrate the results with an example.

Analysis of heat transfer and entropy generation for a thermally developing Brinkman–Brinkman forced convection problem in a rectangular duct with isoflux walls

Heat transfer and entropy generation analysis of the thermally developing forced convection in a porous-saturated duct of rectangular cross-section, with walls maintained at a constant and uniform heat flux, is investigated based on the Brinkman flow model. The classical Galerkin method is used to obtain the fully developed velocity distribution. To solve the thermal energy equation, with the effects of viscous dissipation being included, the Extended Weighted Residuals Method (EWRM) is applied. The local (three dimensional) temperature field is solved by utilizing the Green’s function solution based on the EWRM where symbolic algebra is being used for convenience in presentation. Following the computation of the temperature field, expressions are presented for the local Nusselt number and the bulk temperature as a function of the dimensionless longitudinal coordinate, the aspect ratio, the Darcy number, the viscosity ratio, and the Brinkman number. With the velocity and temperature field being determined, the Second Law (of Thermodynamics) aspect of the problem is also investigated. Approximate closed form solutions are also presented for two limiting cases of MDa values. It is observed that decreasing the aspect ratio and MDa values increases the entropy generation rate.

Reducing problem behavior during care-giving in families of preschool-aged children with developmental disabilities

This study evaluated two variants of a behavioral parent training program known as Stepping Stones Triple P (SSTP) using 74 preschool-aged children with developmental disabilities. Families were randomly allocated to an enhanced parent training intervention that combined parenting skills and care-giving coping skills (SSTP-E), standard parent training intervention alone (SSTP-S) or waitlist control (WL) condition. At post-intervention, both programs were associated with lower levels of observed negative child behavior, reductions in the number of care-giving settings where children displayed problem behavior, and improved parental competence and satisfaction in the parenting role as compared with the waitlist condition. Gains attained at post-intervention were maintained at 1-year follow-up. Both interventions produced significant reductions in child problem behavior, with 67% of children in the SSTP-E and 77% of children in the SSTPS showing clinically reliable change from pre-intervention to follow-up. Parents reported a high level of satisfaction with both interventions.

Studies in search of a suitable experimental insect model for xenodiagnosis of hosts with Chagas' disease: 4 - The reflection of parasite stock in the responsiveness of different vector species to chronic infection with different Trypanosoma cruzi stocks

Previous studies (1982,1987) have emphasized the superiority of sylvatic vector species over domestic species as xenodiagnostic agents in testing hosts with acute or chronic infections by T. cruzi "Y" stock. The present study, which is unique in that it contains data on both infectivity rates produced by the same stock in 11 different vector species and also the reaction of the same vector species to seven different parasite stocks, establishes the general validity of linking efficiency of xenodiagnosis to the biotope of its agent. For example, infectivity rates produced by "São Felipe" stock varied from 82.5% to 98.3% in sylvatic vectors but decreased to 42.5% to 71.3% in domestic species. "Colombiana" stock produced in the same sylvatic vectors infectivity rates ranging from 12.5% to 45%. These shrank to 5%-22.5% in domestic bugs. The functional role of the biotope in the vector-parasite interaction has not been eluddated. But since this phenomenon has been observed to be stable and easy to reproduce, it leads us to believe that the results obtained are valid. Data presented also provide increasing evidence that the infectivity rates exhibited by bugs from xenodiagnosis in chronic hosts, are parasite stock specific. For example, infectivity rates produced by "Berenice", "Y", "FL" and "CL" varied in R. neglectus from 26.3% to 75%; in P. megistus from 56.3% to 83.8%; in T. sordida from 28.8% to 58.8% in T. pseudomaculata from 41.3% to 66.3% and in T. rubrovaria from 48.8% to 85%. Data from xenodiagnosis in the same hosts, carrying acute infections by the same parasite stocks, gave the five sylvatic vectors a positive rating of approximately 100%, thus suggesting that the heavy loads of parasites circulating in the acute hosts obscured the characteristic interspecific differences for the parasite stock. Nonetheless these latter were revealed in the same hosts with chronic infections stimulated by very low numbers of the same parasite stocks. Certain observations here described lead us to speculate as to the possibility of further results from other parasite stocks, allowing the association of the infectivity rates produced in bugs by different parasite stocks with the isoenzymic patterns revealed by these stocks.

Analysis of stock market indices with multidimensional scaling and wavelets

Stock market indices SMIs are important measures of financial and economical performance. Considerable research efforts during the last years demonstrated that these signals have a chaotic nature and require sophisticated mathematical tools for analyzing their characteristics. Classical methods, such as the Fourier transform, reveal considerable limitations in discriminating different periods of time. This paper studies the dynamics of SMI by combining the wavelet transform and the multidimensional scaling MDS . Six continuous wavelets are tested for analyzing the information content of the stock signals. In a first phase, the real Shannon wavelet is adopted for performing the evaluation of the SMI dynamics, while their comparison is visualized by means of the MDS. In a second phase, the other wavelets are also tested, and the corresponding MDS plots are analyzed.