Chalmers Research
Not a member yet
88095 research outputs found
Sort by
Behavior Trees and State Machines in Robotics Applications
No full textAutonomous robots combine skills to form increasingly complex behaviors, called missions. While skills are often programmed at a relatively low abstraction level, their coordination is architecturally separated and often expressed in higher-level languages or frameworks. State machines have been the go-to language to model behavior for decades, but recently, behavior trees have gained attention among roboticists. Originally designed to model autonomous actors in computer games, behavior trees offer an extensible tree-based representation of missions and are claimed to support modular design and code reuse. Although several implementations of behavior trees are in use, little is known about their usage and scope in the real world. How do concepts offered by behavior trees relate to traditional languages, such as state machines? How are concepts in behavior trees and state machines used in actual applications? This paper is a study of the key language concepts in behavior trees as realized in domain-specific languages (DSLs), internal and external DSLs offered as libraries, and their use in open-source robotic applications supported by the Robot Operating System (ROS). We analyze behavior-tree DSLs and compare them to the standard language for behavior models in robotics: state machines. We identify DSLs for both behavior-modeling languages, and we analyze five in-depth. We mine open-source repositories for robotic applications that use the analyzed DSLs and analyze their usage. We identify similarities between behavior trees and state machines in terms of language design and the concepts offered to accommodate the needs of the robotics domain. We observed that the usage of behavior-tree DSLs in open-source projects is increasing rapidly. We observed similar usage patterns at model structure and at code reuse in the behavior-tree and state-machine models within the mined open-source projects. We contribute all extracted models as a dataset, hoping to inspire the community to use and further develop behavior trees, associated tools, and analysis techniques
On Characterization and Optimization of Engineering Surfaces
Get PDFSwedish manufacturing industry in collaboration with academia is exploring innovative ways to manufacture eco-efficient and resource efficient products. Consequently, improving manufacturing efficiency and quality has become the priority for the manufacturing sector to remain competitive in a sustainable way. To achieve this, control and optimization of manufacturing process and product’s performance are necessary. This has led to increase in demand for functional surfaces, which are engineering surfaces tailored to different applications. With new advancements in manufacturing and surface metrology, investigations are steadily progressing towards re-defining quality and meeting dynamic customer demands. In this thesis, surfaces produced by different manufacturing systems are investigated, and methods are proposed to improve specification and optimization.The definition and interpretation of surface roughness vary across the manufacturing industry and academia. It is well known that surface characterization helps to understand the manufacturing process and its influence on surface functional properties such as wear, friction, adhesivity, wettability, fluid retention and aesthetic properties such as gloss. Manufactured surfaces consist of features that are relevant and features that are not of interest. To be able to produce the intended function, it is important to identify and quantify the features of relevance. Use of surface texture parameters helps in quantifying these surface features with respect to type, region, spacing and distribution. Currently, surface parameters Ra or Sa that represent average roughness are widely used in the industry, but they may not provide adequate information on the surface. In this thesis, a general methodology, based on the standard surface parameters and statistical approach, is proposed to improve the specification for surface roughness and identify the combination of significant surface texture parameters that best describe the surface and extract valuable surface information.Surface topography generated by additive, subtractive and formative processes is investigated with the developed research approach. The roughness profile parameters and areal surface parameters defined in ISO, along with power spectral density and scale sensitive fractal analysis, are used for surface characterization and analysis. In this thesis, the application of regression statistics to identify the set of significant surface parameters that improve the specification for surface roughness is shown. These surface parameters are used to discriminate between the surfaces produced by multiple process variables at multiple levels. By analyzing the influence of process variables on the surface topography, the research methodology helps to understand the underlying physical phenomenon and enhance the domain-specific knowledge with respect to surface topography. Subsequently, it helps to interpret processing conditions for process and surface function optimization.The research methods employed in this study are valid and applicable for different manufacturing processes. This thesis can support the guidelines for manufacturing industry focusing on process and functional optimization through surface analysis. With increase in use of machine learning and artificial intelligence in automation, methodologies such as the one proposed in this thesis are vital in exploring and extracting new possibilities in functional surfaces
Hot-carrier generation and strong coupling in metal nanoparticles
Get PDFMetal nanoparticles are an emerging platform for energy applications such as photovoltaics, solar-to-fuel conversion and photocatalysis. Nanoparticles confine the oscillating electromagnetic field of visible light to very small volumes, which allows for efficient absorption and scattering of light in the solar spectral range.The first part of this thesis concerns the formation of so-called hot carriers, which are highly energetic charge carriers that can be generated via the absorption of light and can drive processes relevant for energy applications. The exact mechanisms leading to the formation and transfer of hot carriers are, however, not fully understood, which hinders rational design of nanoparticles for these applications. Here, I have modeled the generation of hot carriers across nanoparticle-molecule junctions by time-dependent density functional theory calculations. I show the importance of energetic alignment between the frontier orbitals, the states in the nanoparticle, and the photon energy for the hot-carrier distribution, leading to a non-monotonic distance dependence.The second part of this thesis focuses on modeling hybrid light-matter states. Hybrid light-matter states can form due the resonant interaction between light and electronic excitations, in a regime of light-matter interaction known as strong coupling. Common approaches for modeling strong coupling are usually limited to highly simplified descriptions of matter. Here, I derive a computationally efficient model based on dipolar coupling. A detailed description of the matter is retained by obtaining polarizabilities of components from time-dependent density functional theory. Finally, I show that the model accurately captures strong coupling behavior in nanoparticle-molecules assemblies
Advances in optimal design and retrofit of chemical processes with uncertain parameters - Applications in design of heat exchanger networks
Get PDFThere is widespread consensus that the omnipresent climate crisis demands humanity to rapidly reduce global greenhouse gas (GHG) emissions.To allow for such a rapid reduction, the industrial sector as a main contributor to GHG emissions needs to take immediate actions. To mitigate GHG emissions from the industrial sector, increasing energy efficiency as well as fuel and feedstock switching, such as increased use of biomass and (green) electricity, are the options which can have most impact in the short- and medium-term.Such mitigation options usually create a need for design of new or redesign of existing processes such as the plant energy systems.The design and operation of industrial plants and processes are usually subject to uncertainty, especially in the process industry. This uncertainty can have different origins, e.g., process parameters such as flow rates or transfer coefficients may vary (uncontrolled) or may not be known exactly.This thesis proposes theoretical and methodological developments for designing and/or redesigning chemical processes which are subject to uncertain operating conditions, with a special focus on heat recovery systems such as heat exchanger networks.In this context, this thesis contributes with theoretical development in the field of deterministic flexibility analysis.More specifically, new approaches are presented to enhance the modelling of the expected uncertainty space, i.e., the space in which the uncertain parameters are expected to vary.Additionally, an approach is presented to perform (deterministic) flexibility analysis in situations when uncertain long-term development such as a switch in feedstocks interferes with operational short-term disturbances.In this context, the thesis presents an industrial case study to i) show the need for such a theoretical development, and ii) illustrate the applicability.Aside of advances in deterministic flexibility analysis, this thesis also explores the possibility to combine valuable designer input (e.g. non-quantifiable knowledge) with the efficiency of mathematical programming when addressing a design under uncertainty problem.More specifically, this thesis proposes to divide the design under uncertainty problem into a design synthesis step which allows direct input from the designer, and several subsequent steps which are summarized in a framework presented in this thesis.The proposed framework combines different approaches from the literature with the theoretical development presented in this thesis, and aims to identify the optimal design specifications which also guarantee that the the final design can operate at all expected operating conditions.The design synthesis step and the framework are decoupled from each other which allows the approach to be applied to large and complex industrial case studies with acceptable computational effort.Usage of the proposed framework is illustrated by means of an industrial case study which presents a design under uncertainty problem
High-level Programming on Low-level Platforms: Two Domain-specific Languages based on Haskell
Get PDFIn this thesis, we describe our research on how to program low-level platforms with high-level languages.As an example, consider applications that run on microcontrollers. Such applicationsmay need to specify precise temporal behavior, carefully manage power usage, and handle cryptographic keys.Low-level platforms are programmed using low-level languages such as C/C++, where the lack of expressiveness canlead to error-prone code.We investigate whether we can use high-level languages to program these platforms, by embedding domain-specific languagesin a host language, Haskell. A high-level language offers better expressivity and shields the developer from low-leveldetails, yielding code that more concretely describes what the application is supposed to do. Furthermore, a richer runtimesystem could ease the burden of e.g. memory management and scheduling of coroutines.The papers in this thesis indicate that it is possible to program these devices using a high-level language. We developtwo domain-specific languages, Scoria and HasTEE. Scoria is evaluated on NRF52 microcontrollers, where we run applicationsthat require precise, temporal behavior and perform I/O. HasTEE is evaluated on machines whose processor has supportfor Intel Software Guard Extension and shows that the type system of Haskell can be used to automatically partition a Haskellapplication and run it in a trusted execution environment
Enhancing the feasibility of sustainable biofuel production - The role of inorganic content and potential removal strategies from fast pyrolysis bio-oil based on residues
Get PDFBiofuels made from residual non-edible biomass, like agricultural and forestry residues, are crucial for the transition to a biobased circular economy. These biofuels reduce dependence on non-renewable resources, emit less greenhouse gases than fossil fuels and subsequently contribute to meeting the climate goals set by the Paris Agreement. Additionally, they reduce byproducts by valorizing materials that would otherwise be used as lower value product such as heat. Fast pyrolysis, which is a process at elevated temperatures in the absence of oxygen, converts solid biomass residues into a liquid within seconds. The liquid has a chemical structure that makes it a potential resource to be further upgraded into a biofuel. Due to the high water and oxygen content of the pyrolysis oil, as well as its high acidity, a catalytic process for removing oxygen with hydrogen can be used. However, the presence of inorganic elements poses a risk of deactivating these catalysts during the upgrading process. This limits the use of some biomass resources which naturally contain a low but varying concentration of alkali and alkali earth metals, transition metals and non-metals. The focus of this licentiate thesis is the inorganic content of fast pyrolysis bio-oil produced from residual, non-edible biomass. In the first study, removal of inorganic elements from pyrolysis oil was studied using sorbent materials. The process removed 85-98 % of phosphorus (P), calcium (Ca), iron (Fe), potassium (K), and magnesium (Mg) at 30 \ub0C, ambient pressure and 4 h residence time. It was found that Lewis acid sites on aluminum oxide material is likely key in the removal of P, while strong Br\uf8nsted acid sites on an acidic zeolite and strongly acidic ion-exchange resins are vital for the adsorption of Ca, Fe, K, and Mg. However, the inorganic elements in pyrolysis oil are not only in ionic forms. The organic content in the oil was studied before and after the sorbent removal using 2D gas chromatography, and showed negligible effects on contents of alcohols, phenols, acids, furans, sugars, and ketones/aldehydes, likely due to the mild conditions used. In the second study, water extraction of pyrolysis oil was examined to study the distribution and solubility of the inorganic elements and their forms present in pyrolysis oil. It was found that K and Mg were to a greater extent present in a water-soluble form, compared to Ca, Fe, and P, which to a greater extent were suggested to be part of ash particles or possibly soluble in the organic fraction. This effect was more pronounced when investigating a pyrolysis oil with higher inorganic content. During the water extraction, the mass yield of carbon was 50 % indicating a significant amount in the feed oil is polar. Using 31P nuclear magnetic resonance analysis, it was shown that it consisted to a large part of aliphatic hydroxyls.The removal of the inorganic content from fast pyrolysis oil is crucial for the development of cost-efficient production of biofuels. By increasing the available biomass feedstocks there is a potential of increasing its cost-competitiveness to fossil fuels, which is especially important for parts of the transport sector for which electrification is impractical. The utilization of all sources that can contribute to sustainable advanced biofuels represents a significant step towards our common goal of transitioning to a fossil-free economy and climate change mitigation
How, not if, is the question mycologists should be asking about DNA-based typification
Get PDFFungal metabarcoding of substrates such as soil, wood, and water is uncovering an unprecedented number of fungal species that do not seem to produce tangible morphological structures and that defy our best attempts at cultivation, thus falling outside the scope of the International Code of Nomenclature for algae, fungi, and plants. The present study uses the new, ninth release of the species hypotheses of the UNITE database to show that species discovery through environmental sequencing vastly outpaces traditional, Sanger sequencing-based efforts in a strongly increasing trend over the last five years. Our findings chal-lenge the present stance of some in the mycological community - that the current situation is satisfactory and that no change is needed to "the code" - and suggest that we should be discussing not whether to allow DNA-based descriptions (typifications) of species and by extension higher ranks of fungi, but what the precise requirements for such DNA-based typifications should be. We submit a tentative list of such criteria for further discussion. The present authors hope for a revitalized and deepened discussion on DNA-based typification, because to us it seems harmful and counter-productive to intentionally deny the overwhelming majority of extant fungi a formal standing under the International Code of Nomenclature for algae, fungi, and plants
Disk Wind Feedback from High-mass Protostars. II. The Evolutionary Sequence
Get PDFStar formation is ubiquitously associated with the ejection of accretion-powered outflows that carve bipolar cavities through the infalling envelope. This feedback is expected to be important for regulating the efficiency of star formation from a natal prestellar core. These low-extinction outflow cavities greatly affect the appearance of a protostar by allowing the escape of shorter-wavelength photons. Doppler-shifted CO line emission from outflows is also often the most prominent manifestation of deeply embedded early-stage star formation. Here, we present 3D magnetohydrodynamic simulations of a disk wind outflow from a protostar forming from an initially 60 M ⊙ core embedded in a high-pressure environment typical of massive star-forming regions. We simulate the growth of the protostar from m * = 1 M ⊙ to 26 M ⊙ over a period of ∼100,000 yr. The outflow quickly excavates a cavity with a half opening angle of ∼10\ub0 through the core. This angle remains relatively constant until the star reaches 4 M ⊙. It then grows steadily in time, reaching a value of ∼50\ub0 by the end of the simulation. We estimate a lower limit to the star formation efficiency (SFE) of 0.43. However, accounting for continued accretion from a massive disk and residual infall envelope, we estimate that the final SFE may be as high as ∼0.7. We examine observable properties of the outflow, especially the evolution of the cavity\u27s opening angle, total mass, and momentum flux, and the velocity distributions of the outflowing gas, and compare with the massive protostars G35.20-0.74N and G339.88-1.26 observed by the Atacama Large Millimeter/submillimeter Array (ALMA), yielding constraints on their intrinsic properties
A Novel Hybrid Model for the Evaluation of Industry 4.0 Technologies\u27 Applicability in Logistics Centers
Get PDFThe application of Industry 4.0 (I4.0) in the field of logistics leads to the emergence and development of the concept of logistics 4.0. Many I4.0 technologies have been applied in the field of logistics. The goal of this research is to analyze the applicability of nine key I4.0 technologies in logistics centers (LC). For this purpose, an integrated MEREC (MEthod based on the Removal Effects of Criteria) - fuzzy MARCOS (Measurement of Alternatives and Ranking according to COmpromise Solution) model was developed. The applicability of nine I4.0 technologies was evaluated based on 15 subcriteria within three main groups of criteria, namely, technological, social and political, and economic and operative. Using the MEREC method, the weight values of the criteria and subcriteria were determined, while the technologies were ranked using the fuzzy MARCOS method. Based on the results obtained by applying this integrated MCDM (multicriteria decision-making) model, CC was identified as the best alternative, i.e., the technology that is most applicable in logistics centers, followed by IoT and big data. An analysis of the sensitivity of the obtained results to the change in the importance of the criteria was carried out, which shows certain changes in the ranking when the importance of the most important criterion changes