51 research outputs found
Centrifugal forming and mechanical properties of silicone-based elastomers for soft robotic actuators
No full textThis thesis describes the centrifugal forming and resulting mechanical properties of silicone-based elastomers for the manufacture of soft robotic actuators. This process is effective at removing bubbles that get entrapped within 3D-printed, enclosed molds. Conventional methods for rapid prototyping of soft robotic actuators to remove entrapped bubbles typically involve degassing under vacuum, with open-faced molds that limit the layout of formed parts to raised 2D geometries. As the functionality and complexity of soft robots increase, there is a need to mold complete 3D structures with controlled thicknesses or curvatures on multiples surfaces. In addition, characterization of the mechanical properties of common elastomers for these soft robots has lagged the development of new designs. As such, relationships between resulting material properties and processing parameters are virtually non-existent. One of the goals of this thesis is to provide guidelines and physical insights to relate the design, processing conditions, and resulting properties of soft robotic components to each other. Centrifugal forming with accelerations on the order of 100 g’s is capable of forming bubble-free, true 3D components for soft robotic actuators, and resulting demonstrations in this work include an aquatic locomotor, soft gripper, and an actuator that straightens when pressurized. Finally, this work shows that the measured mechanical properties of 3D geometries fabricated within enclosed molds through centrifugal forming possess comparable mechanical properties to vacuumed materials formed from open-faced molds with raised 2D features.M.S.Includes bibliographical referencesby Parth Kulkarn
Synthetic Data Generation for Bridging Sim2Real Gap in a Production Environment
No full textSynthetic data is being used lately for training deep neural networks in
computer vision applications such as object detection, object segmentation and
6D object pose estimation. Domain randomization hereby plays an important role
in reducing the simulation to reality gap. However, this generalization might
not be effective in specialized domains like a production environment involving
complex assemblies. Either the individual parts, trained with synthetic images,
are integrated in much larger assemblies making them indistinguishable from
their counterparts and result in false positives or are partially occluded just
enough to give rise to false negatives. Domain knowledge is vital in these
cases and if conceived effectively while generating synthetic data, can show a
considerable improvement in bridging the simulation to reality gap. This paper
focuses on synthetic data generation procedures for parts and assemblies used
in a production environment. The basic procedures for synthetic data generation
and their various combinations are evaluated and compared on images captured in
a production environment, where results show up to 15% improvement using
combinations of basic procedures. Reducing the simulation to reality gap in
this way can aid to utilize the true potential of robot assisted production
using artificial intelligence.Comment: 17 pages, 9 figures, LaTeX; typos corrected; has not been presented
in any conference or published in journa
Synthetic data generation procedures for domain-specific environments in manufacturing
No full text16681679Synthetic data is being used lately for training deep neural networks in computer vision applications such as object detection, object segmentation and 6D object pose estimation. Domain randomization hereby plays an important role in reducing the simulation to reality gap. However, this generalization might not always be effective in specialized domains like manufacturing that involve complex assemblies. Individual parts are integrated in much larger assemblies making them indistinguishable from their counterparts. Moreover, individual parts are often partially occluded in the scene. These situations give rise to wrong detections. Target domain knowledge is vital in these cases and if conceived effectively while generating synthetic data, can show a considerable improvement in bridging the simulation to reality gap. This paper validates synthetic data generation procedures through practical experimentation ensuring that experiments are both comprehensive and reproducible. After combining domain randomization and domain adaptation procedures for parts and assemblies used in manufacturing the model performance improves by up to 15% than the state-of-the-art domain randomization techniques. Reducing the simulation to reality gap in this way can unlock the true potential of robot-assisted production using artificial intelligence.25
Synthetic data generation procedures for domain-specific environments in manufacturing
No full textSynthetic data is being used lately for training deep neural networks in computer vision applications such as object detection, object segmentation and 6D object pose estimation. Domain randomization hereby plays an important role in reducing the simulation to reality gap. However, this generalization might not always be effective in specialized domains like manufacturing that involve complex assemblies. Individual parts are integrated in much larger assemblies making them indistinguishable from their counterparts. Moreover, individual parts are often partially occluded in the scene. These situations give rise to wrong detections. Target domain knowledge is vital in these cases and if conceived effectively while generating synthetic data, can show a considerable improvement in bridging the simulation to reality gap. This paper validates synthetic data generation procedures through practical experimentation ensuring that experiments are both comprehensive and reproducible. After combining domain randomization and domain adaptation procedures for parts and assemblies used in manufacturing the model performance improves by up to 15% than the state-of-the-art domain randomization techniques. Reducing the simulation to reality gap in this way can unlock the true potential of robot-assisted production using artificial intelligence
Synthetic data generation for bridging Sim2Real gap in a production environment
No full textSynthetic data is being used lately for training deep neural networks in computer vision applications such as object detection, object segmentation and 6D object pose estimation. Domain randomization hereby plays an important role in reducing the simulation to reality gap. However, this generalization might not be effective in specialized domains like a production environment involving complex assemblies. Either the individual parts, trained with synthetic images, are integrated in much larger assemblies making them indistinguishable from their counterparts and result in false positives or are partially occluded just enough to give rise to false negatives. Domain knowledge is vital in these cases and if conceived effectively while generating synthetic data, can show a considerable improvement in bridging the simulation to reality gap. This paper focuses on synthetic data generation procedures for parts and assemblies used in a production environment. The basic procedures for synthetic data generation and their various combinations are evaluated and compared on images captured in a production environment, where results show up to 15% improvement using combinations of basic procedures. Reducing the simulation to reality gap in this way can aid to utilize the true potential of robot assisted production using artificial intelligence
Synthetic Data Generation for Bridging Sim2Real Gap in a Production Environment
No full textSynthetic data is being used lately for training deep neural networks in computer vision applications such as object detection, object segmentation and 6D object pose estimation. Domain randomization hereby plays an important role in reducing the simulation to reality gap. However, this generalization might not be effectie in specialized domains like a production environment involving complex assemblies. Either the individual parts, trained with synthetic images, are integrated in much larger assemblies making them indistinguishable from their counterparts and result infalse positives or are partially occluded just enough to give rise to false negatives. Domain knowledge is vital in these cases and if conceived effectiely while generating synthetic data, can show a considerable improvement in bridging the simulation to reality gap. This paper focuses on synthetic data generation procedures for parts and assemblies used in a production environment.The basic procedures for synthetic data generation and their various combinations are evaluatedand compared on images captured in a production environment, where results show up to 15% improvement using combinations of basic procedures. Reducing the simulation to reality gap in this way can aid to utilize the true potential of robot assisted production using artificial intelligence
Characterization and optimization of UAV power system for aerial and submersible multi-medium multirotor vehicle
No full textEven as an emerging technology, Unmanned Aerial Vehicles (UAVs) have had a tremendous impact on the world. From the way wars are fought, to the way we take selfies, drones are well on their way to revolutionizing our daily lives. One of the most innovative applications of these vehicles in the Naviator submersible-UAV. This unique multirotor is capable of aerial flight and underwater operations with seamless Air-Water transitions. In this thesis, the power system of a multirotor UAS is characterized using standard performance models with the goal of designing and optimizing the systems of a new Naviator V5 prototype. Test beds were created to collect data on BLDC motors and propellers and their performance was assessed in air and water. Theoretical models using BEM theory and the 3-constant motor model were validated for their accuracy. Experiments found that RC air propellers are similarly efficient in air and water and BLDC motor performance is partially diminished due to the higher viscosity of water. The effects of input voltage, throttle, Kv rating, and motor size were also evaluated using motor torque curves. Using this data, an optimal power system for the Naviator V5 prototype was designed, tested, and evaluated.M.S.T.Includes bibliographical referencesby Parth V. Son
Evacuation of Kobylnice village in extraordinary event
No full textCílem této práce je sestavení evakuačního plánu pro obec Kobylnice. Teoretická část přibližuje problematiku mimořádných událostí a možných následků, zpracovává dostupnou literaturu a uvádí zákony a vyhlášky spojené s evakuačním plánováním. V praktické části se autor zmiňuje o programech, které jsou použitelné při mimořádných a krizových situacích a sestavuje plán evakuace.The aim of this work is evacuation plan setup for Kobylnice Corporation. The theoretical parth explains issue of extraordinary events and possible consequences, processes available information and shows the laws and public notices are being used in evacuative planning. In the practical parth author presents programs, which are applicability to extraordinary events and crisis situations. She draws a plan of evacuation.Institut bezpečnostních technologiíobhájen
Foundations of the law on industrial organisations in Russia and the former republics of the USSR: 1985-1990
No full textThe industrial organisations introduced into the law of the USSR from 1987, and thereafter
into the law of the former republics, developed upon a foundation that was rooted in Soviet
law and was constructed during the period from 1985 to mid-1990.
While this study focuses on the industrial economy, certain aspects of the agrarian economy,
and in particular the early history and structure of the collective farm, are considered where
appropriate.
The thesis presents an entirely new understanding both of the nature of these developments
and of the significance of the law on ownership. The foundations of the law on industrial
organisations are conceptualised within specific heuristic models which are elaborated in an
attempt to consolidate and highlight the key steps in this history. It is argued that Soviet law
did not contain a concept of the "generic owner" or a developed understanding of the
ownership of a juridical person, in particular by multiple owners holding "ownership
interests" of that juridical person; and that their absence critically impaired a rational and
coherent structure for the foundations of the law on industrial organisations both within the
Stalin economic settlement and the new economic constitution of 1990
Electrochemical screening and surface analysis of environmentally friendly corrosion inhibitors for aerospace aluminium alloy
No full textFor the past few decades, the aerospace industry has been actively looking for measures to tackle the problem of localized corrosion of aluminium alloy 2024-T3. One such measure was the use of hexavalent chromium salts as corrosion inhibitors. However, recent studies have reflected the environmental concerns and health impacts associated with these compounds. Since then, there has been a quest for alternatives that can be used as corrosion inhibitors for AA2024-T3. This thesis is aimed at studying these compounds by testing their inhibitive properties. Selecting an inhibitor from a set of compounds and recommending the best one is a crucial process. The research focuses on the screening of these inhibitors through various electrochemical techniques such as Potentiodynamic Polarisation(PDP), Linear Polarisation Resistance(LPR), and Electrochemical Impedance Spectroscopy(EIS). The electrochemical experiments during the first 24-hour exposure of AA2024-T3 to inhibitors exhibited that 2,5-Dimercapto-1,3,4-thiadiazole acts as a corrosion accelerator, while Na-Mercaptoacetate and Mercaptobenzimidazole showed the most inhibitive performance among the tested organic compounds. On monitoring their nature during 24 hours, it was observed that they show stable polarisation resistance values after 6 hours. Among the inorganics, Cerium compounds were seen to have better inhibitive properties than the Lithiumcompounds. Additionally, surface studies such as Fourier Transform Infrared Spectroscopy (FTIR) revealed the presence of thiol/mercaptan and carboxyl groups on samples exposed toNa-Mercaptoacetate. X-ray Photoelectron Spectroscopy (XPS) showed oxides of Cerium on samples exposed to Cerium Nitrate. Finally, the correlations between the inhibition efficiencies calculated from different electrochemical testing methods were illustrated using Pearson’s correlation method.Materials Science and Engineerin
- …
