Classification of Assembly Operations Using Recurrent Neural Networks

Visual sensor data of manual assembly operations offers rich information that can be extracted in order to analyze and digitalize the assembly. The worker’s interaction with tools and objects, as well as the spatial–temporal nature of assembly operations, makes the recognition and classification of assembly operations a complex task. Therefore, classical methods of computer vision do not provide a sufficient solution. This paper presents a recurrent neural network for the classification of manual assembly operations using visual sensor data and addresses the question as to what extent such a solution is feasible in terms of robustness and reliability. Since complex assembly operations are a combination of basic movements, four main assembly operations of the Methods Time-Measurement base operations are classified using a machine learning approach. A dataset of these four assembly operations, reach, grasp, move and release, containing RGB-, infrared-, and depth-data is used. A Convolutional Neural Network—Long Short Term Memory architecture is investigated regarding its applicability due to the spatial–temporal nature of the data.

Strain Analysis of Ti6Al4V Titanium Alloy Samples Using Digital Image Correlation

Digital image correlation (DIC) is a non-contact optical method that allows measuring displacements on a plane used to determine the strains caused by external loads of a structural element (mechanical or thermal). Currently, digital image correlation is a widely used experimental technique to assess the mechanical behavior of materials, in particular cracking characteristics and destruction methods of various structural elements. In this paper, the DIC method is applied to determine local strains of titanium alloy Ti6Al4V specimen. The samples used in the tests were made with two different technologies: (a) from a drawn bar by machining process; and (b) by the additive manufacturing method Direct Metal Laser Sintering (DMLS). The aim of the paper is to present the mechanical properties test results of the Ti6Al4V titanium alloy produced by the DMLS additive manufacturing under static loads using the digital image correlation method. As a result of the tests carried out on the drawn bar specimens, it was concluded that the change in the measurement base affects the difference in the Young’s E modulus value in the range from 89.2 to 103.8 GPa. However, for samples formed using the DMLS method, the change in the Young’s modulus value was from 112.9 to 115.3 GPa for the same measurement base.

THE ALGORITHM OF COMPLEX DIAGNOSTICS OF TECHNICAL CONDITION OF BUILDING STRUCTURES ON THERMOGRAMS ANALYSIS

the algorithm of complex diagnostics of a technical condition of building constructions and structures on the analysis of thermograms is offered. An attempt is made to systematize and logically build a sequence of stages of energy inspection of heat-consuming and heat-generating construction projects and their structural elements. The presented algorithm includes the main stages performed in such surveys. It differs from the known technologies of energy monitoring of buildings and structures by the presence of two additional stages, which are an integral part of any energy survey: determination of the relative surface humidity of building structures and the actual mass flow of coolant through the heating device for the analysis of thermal images. The implementation of energy (thermal imaging) and technical inspections according to the presented structural scheme requires not only the presence of appropriate instrumentation and measurement base, but also the involvement of highly qualified specialists (civil engineers, thermal power engineers, thermal physicists, operators-thermographers, etc.).

Development of Quantum Private Queries Protocol on Collective-Dephasing Noise Channel

Quantum private queries can commonly protect important information in a good many of domains, such as finance, business, military, which use quantum effects to achieve unprecedented classical private queries. However, quantum state can be easily affected by environmental noise, which affects the actual effect of quantum private queries. This paper developed a new quantum private query protocol based on four qubits logical Bell state to resist the collective-dephasing noise. The symmetric private information retrieval problem, which is the most influential problem in the process of quantum private query, was solved well by quantum oblivious transfer. It introduces the construction of four qubits logical Bell state. The quantum private query protocol innovates the quantum key distribution process by using the four qubits logical Bell state as the measurement base to measure the logical qubits, and ensures the function of quantum oblivious transmission. The protocol cannot only resist the noise influence of the communication process, but also ensure the security of both sides of the communication.

Problems in estimate of parts geometrical accuracy

The paper reports the analysis of the method for parts geometrical accuracy definition which includes the following stages: part surfaces measurement, a representation of measurement results and their relative position, a definition without errors count and a definition of surface errors. There are shown drawbacks of stages enumerated such as a neglect of measurement base surface errors, an ambiguity of measurement base positions regarding surfaces under control, use of bases of error account and criteria for errors estimate irrespective of surface operation functions. A multi-gradation of the method for the estimate of a part error resulting in the accumulation of total error is shown.