scholarly journals OPTIMAL POSITIONING METHODS OF INTEGRAL DEFORMATION SENSORS – EXPERT KNOWLEDGE VERSUS MATHEMATICAL OPTIMIZATION

2021 ◽  
Vol 2021 (3) ◽  
pp. 4628-4635
Author(s):  
Ch. Brecher ◽  
◽  
R. Herzog ◽  
A. Naumann ◽  
R. Spierling ◽  
...  
Keyword(s):  
Machine Tool ◽  
Domain Knowledge ◽  
Expert Knowledge ◽  
Kinematic Model ◽  
Mathematical Optimization ◽  
Elastic Behavior ◽  
Work Piece ◽  
Mathematical Methods ◽  
Optimal Position ◽  
Structural Deformations

Up to 75 % of the overall work piece error can be caused by the thermo-elastic behavior of the machine tool. Therefore, correction methods based on machine-integrated sensors were intensively researched during the last years, in order to determine the error of the Tool Center Point (TCP) parallel to the process. One of these methods includes the integral deformation sensor (IDS), which detects the deformation along the length of a structural component of the machine. The error of the TCP is modelled based on the measured structural deformations, a mechanical model of the structural parts and a kinematic model of the machine tool. Currently, the sensor setup for specific machines is usually defined by an expert with the help of his or her domain knowledge. There are existing mathematical methods for optimal sensor positioning. The aim of this work is the evaluation of the expert positioning versus the mathematical methods. The parameters to be varied are the lengths and positions of the IDS. Criteria for the evaluation are the achievable accuracy of the TCP error prediction and the sensitivity to small variations of the optimal position, as they might occur during the installation.

scholarly journals Expert-enhanced machine learning for cardiac arrhythmia classification

PLoS ONE ◽  
2021 ◽  
Vol 16 (12) ◽  
pp. e0261571
Author(s):  
Sebastian Sager ◽  
Felix Bernhardt ◽  
Florian Kehrle ◽  
Maximilian Merkert ◽  
Andreas Potschka ◽  
...  
Keyword(s):  
Machine Learning ◽  
Atrial Flutter ◽  
Cardiac Arrhythmia ◽  
Domain Knowledge ◽  
Expert Knowledge ◽  
Mathematical Optimization ◽  
Regression Problem ◽  
Additional Advantage ◽  
Inverse Simulation

We propose a new method for the classification task of distinguishing atrial fibrillation (AFib) from regular atrial tachycardias including atrial flutter (AFlu) based on a surface electrocardiogram (ECG). Recently, many approaches for an automatic classification of cardiac arrhythmia were proposed and to our knowledge none of them can distinguish between these two. We discuss reasons why deep learning may not yield satisfactory results for this task. We generate new and clinically interpretable features using mathematical optimization for subsequent use within a machine learning (ML) model. These features are generated from the same input data by solving an additional regression problem with complicated combinatorial substructures. The resultant can be seen as a novel machine learning model that incorporates expert knowledge on the pathophysiology of atrial flutter. Our approach achieves an unprecedented accuracy of 82.84% and an area under the receiver operating characteristic (ROC) curve of 0.9, which classifies as “excellent” according to the classification indicator of diagnostic tests. One additional advantage of our approach is the inherent interpretability of the classification results. Our features give insight into a possibly occurring multilevel atrioventricular blocking mechanism, which may improve treatment decisions beyond the classification itself. Our research ideally complements existing textbook cardiac arrhythmia classification methods, which cannot provide a classification for the important case of AFib↔AFlu. The main contribution is the successful use of a novel mathematical model for multilevel atrioventricular block and optimization-driven inverse simulation to enhance machine learning for classification of the arguably most difficult cases in cardiac arrhythmia. A tailored Branch-and-Bound algorithm was implemented for the domain knowledge part, while standard algorithms such as Adam could be used for training.

scholarly journals A Kinematic Model to Compensate the Structural Deformations in Machine Tools Using Fibre Bragg Grating (FBG) Sensors

2016 ◽  
Author(s):  
Francesco Aggogeri ◽  
Alberto Borboni ◽  
Rodolfo Faglia ◽  
Angelo Merlo ◽  
Nicola Pellegrini
Keyword(s):  
Machine Tool ◽  
Real Time ◽  
Kinematic Model ◽  
Experimental Tests ◽  
Bragg Grating ◽  
Fibre Bragg Grating ◽  
Machining Performance ◽  
Novel Approach ◽  
Structural Deformations ◽  
Fbg Sensors

Structural deformations are one of the most significant factor that affects machine tool (MT) positioning accuracy. These induced errors are complex to be represented by a model, nevertheless they need to be evaluated and predicted in order to increase the machining performance. This paper presents a novel approach to calibrate a machine tool in real-time, analyzing the thermo-mechanical errors through Fibre Bragg Grating (FBG) sensors embedded in the MT frame. The proposed configuration consists of an adaptronic structure of passive materials, Carbon Fibre Reinforced Polymers (CFRP), equipped by FBG sensors that are able to measure in real-time the deformed conditions of the frame. By using a proper thermo-mechanical kinematic model, the displacement of the end effector may be predicted and corrected when it is subjected to external undesired factors. By starting from a set of FE simulations to develop a model able to describe the MT structure stresses, a prototype has been fabricated and tested. The scope was to compare the numerical model with the experimental tests using FBG sensors. The experimental campaign has been performed varying the structure temperature over time and measuring the tool tip point (TTP) positions. The obtained results showed a substantial matching between the real and the predicted position of TTP confirming the effectiveness of the proposed calibration system.

scholarly journals Templated Text Synthesis for Expert-Guided Multi-Label Extraction from Radiology Reports

2021 ◽  
Vol 3 (2) ◽  
pp. 299-317
Author(s):  
Patrick Schrempf ◽  
Hannah Watson ◽  
Eunsoo Park ◽  
Maciej Pajak ◽  
Hamish MacKinnon ◽  
...  
Keyword(s):  
Domain Knowledge ◽  
Expert Knowledge ◽  
Medical Image Analysis ◽  
Imaging Data ◽  
Data Synthesis ◽  
Domain Specific ◽  
Radiology Reports ◽  
Training Examples ◽  
Analysis Models ◽  
Difficult Cases

Training medical image analysis models traditionally requires large amounts of expertly annotated imaging data which is time-consuming and expensive to obtain. One solution is to automatically extract scan-level labels from radiology reports. Previously, we showed that, by extending BERT with a per-label attention mechanism, we can train a single model to perform automatic extraction of many labels in parallel. However, if we rely on pure data-driven learning, the model sometimes fails to learn critical features or learns the correct answer via simplistic heuristics (e.g., that “likely” indicates positivity), and thus fails to generalise to rarer cases which have not been learned or where the heuristics break down (e.g., “likely represents prominent VR space or lacunar infarct” which indicates uncertainty over two differential diagnoses). In this work, we propose template creation for data synthesis, which enables us to inject expert knowledge about unseen entities from medical ontologies, and to teach the model rules on how to label difficult cases, by producing relevant training examples. Using this technique alongside domain-specific pre-training for our underlying BERT architecture i.e., PubMedBERT, we improve F1 micro from 0.903 to 0.939 and F1 macro from 0.512 to 0.737 on an independent test set for 33 labels in head CT reports for stroke patients. Our methodology offers a practical way to combine domain knowledge with machine learning for text classification tasks.

The Determination of the Three-Point Support Design of Machine Tool Based on iSIGHT

2014 ◽  
Vol 607 ◽  
pp. 342-345
Author(s):  
Sheng Hui Zhao ◽  
Xiao Chuang Zhu ◽  
Da Wei Zhang
Keyword(s):  
Machine Tool ◽  
High Precision ◽  
Multidisciplinary Optimization ◽  
Joint Surface ◽  
Objective Functions ◽  
Multi Objective Optimization ◽  
Support Design ◽  
Optimal Position ◽  
Point Support

In order to meet the requirements of high-precision machine tool, it has been an important factor to select an appropriate way to support the bed. By building a multidisciplinary optimization (MDO) process based on iSIGHT, this article select the deformation difference of the guides and the deformation difference of the joint surface between column and bed of the machine tool as the objective functions, and then conduct a multi-objective optimization (MOO) of the positional parameters of the three-point support. Eventually the optimization result is given and the optimal position of the three-point support is determined.

Design of a Force Feedback Chatter Control System

1972 ◽  
Vol 94 (1) ◽  
pp. 5-10 ◽  
Author(s):  
C. Nachtigal
Keyword(s):  
Machine Tool ◽  
Frequency Domain ◽  
Transient Response ◽  
Force Feedback ◽  
Design Procedure ◽  
Experimental Tests ◽  
Work Piece ◽  
Regenerative Chatter ◽  
Chatter Control ◽  
The Stability

The analysis of machine tool chatter from frequency domain considerations is generally accepted as a valid representation of the regenerative chatter phenomenon. However, active control of regenerative chatter is still in its embryonic stage. It was established in reference [2] that a measurement of the cutting force could be effectively used in conjunction with a controller and a tool position servo system to increase the stability of an engine lathe and to improve its transient response. This paper presents the design basis for such a system, including both analytical and experimental considerations. The design procedure stems from a real part stability criterion based on the work by Merritt [1]. Because of the unknown variability in the dynamics of a machine tool system, the controller parameters were chosen to accomodate some mismatch between structure and tool servo dynamics. Experimental tests to determine the stability zone of the controlled machine tool system qualitatively confirmed the analytical design results. The experimental results were consistent in that the transient response tests confirmed the frequency domain stability tests. It was also demonstrated experimentally that the equivalent static stiffness of a flexible work-piece system could be substantially increased.

A kinematic approach for error modelling in drilling

2019 ◽  
Vol 36 (4) ◽  
pp. 1364-1383 ◽  
Author(s):  
Wilma Polini ◽  
Andrea Corrado
Keyword(s):  
Machine Tool ◽  
Reference Frame ◽  
Kinematic Model ◽  
Model Parameters ◽  
Geometric Errors ◽  
Volumetric Error ◽  
Machined Surface ◽  
Content Type ◽  
Practical Applications ◽  
Source Errors

Purpose The purpose of this paper is to model how geometric errors of a machined surface (or manufacturing errors) are related to locators’ error, workpiece form error and machine tool volumetric error. A kinematic model is presented that puts into relationship the locator error, the workpiece form deviations and the machine tool volumetric error. Design/methodology/approach The paper presents a general and systematic approach for geometric error modelling in drilling because of the geometric errors of locators positioning, of workpiece datum surface and of machine tool. The model can be implemented in four steps: (1) calculation of the deviation in the workpiece reference frame because of deviations of locator positions; (2) evaluation of the deviation in the workpiece reference frame owing to form deviations in the datum surfaces of the workpiece; (3) formulation of the volumetric error of the machine tool; and (4) combination of those three models. Findings The advantage of this approach lies in that it enables the source errors affecting the drilling accuracy to be explicitly separated, thereby providing designers and/or field engineers with an informative guideline for accuracy improvement through suitable measures, i.e. component tolerancing in design, machining and so on. Two typical drilling operations are taken as examples to illustrate the generality and effectiveness of this approach. Research limitations/implications Some source errors, such as the dynamic behaviour of the machine tool, are not taken into consideration, which will be modelled in practical applications. Practical implications The proposed kinematic model may be set by means of experimental tests, concerning the industrial specific application, to identify the values of the model parameters, such as standard deviation of the machine tool axes positioning and rotational errors. Then, it may be easily used to foresee the location deviation of a single or a pattern of holes. Originality/value The approaches present in the literature aim to model only one or at most two sources of machining error, such as fixturing, machine tool or workpiece datum. This paper goes beyond the state of the art because it considers the locator errors together with the form deviation on the datum surface into contact with the locators and, then, the volumetric error of the machine tool.

Programming Language Support for Implementing Machine Learning Algorithms

2020 ◽  
pp. 402-421
Author(s):  
Anitha Elavarasi S. ◽  
Jayanthi J.
Keyword(s):  
Machine Learning ◽  
Programming Languages ◽  
Programming Language ◽  
Domain Knowledge ◽  
Learning Algorithm ◽  
Mathematical Optimization ◽  
Machine Learning Algorithms ◽  
Application Development ◽  
Machine Learning Applications ◽  
And Control

Machine learning provides the system to automatically learn without human intervention and improve their performance with the help of previous experience. It can access the data and use it for learning by itself. Even though many algorithms are developed to solve machine learning issues, it is difficult to handle all kinds of inputs data in-order to arrive at accurate decisions. The domain knowledge of statistical science, probability, logic, mathematical optimization, reinforcement learning, and control theory plays a major role in developing machine learning based algorithms. The key consideration in selecting a suitable programming language for implementing machine learning algorithm includes performance, concurrence, application development, learning curve. This chapter deals with few of the top programming languages used for developing machine learning applications. They are Python, R, and Java. Top three programming languages preferred by data scientist are (1) Python more than 57%, (2) R more than 31%, and (3) Java used by 17% of the data scientists.

Development and Analysis of an Enhanced Multi-Expert Knowledge Integration System for Designing Context-Aware Ubiquitous Learning Contents

2018 ◽  
Vol 16 (4) ◽  
pp. 31-53 ◽  
Author(s):  
Gwo-Haur Hwang ◽  
Beyin Chen ◽  
Shiau-Huei Huang
Keyword(s):  
Knowledge Acquisition ◽  
Real World ◽  
Domain Knowledge ◽  
Knowledge Integration ◽  
Expert Knowledge ◽  
Context Aware ◽  
Ubiquitous Learning ◽  
Integration System ◽  
Degree Of Acceptance ◽  
Multiple Experts

This article describes how in context-aware ubiquitous learning environments, teachers must plan a theme and design learning contents to provide complete knowledge for students. Knowledge acquisition, which is an approach for helping people represent and organize domain knowledge, has been recognized as a potential way of guiding teachers to develop real-world context-related learning contents. However, previous studies failed to address the issue that the learning contents provided by multiple experts or teachers might be redundant or inconsistent; moreover, it is difficult to use the traditional knowledge acquisition method to fully describe the complex real-world contexts and the learning contents. Therefore, in this article, a multi-expert knowledge integration system with an enhanced knowledge representation approach and Delphi method has been developed. From the experimental results, it is found that the teachers involved had a high degree of acceptance of the system. They believe that it can unify the knowledge of many teachers.

scholarly journals A novel six-legged walking machine tool for in-situ operations

2020 ◽  
Vol 15 (3) ◽  
pp. 351-364
Author(s):  
Jimu Liu ◽  
Yuan Tian ◽  
Feng Gao
Keyword(s):  
Machine Tool ◽  
Motion Planning ◽  
Kinematic Model ◽  
The Body ◽  
Body Motion ◽  
Walking Distance ◽  
Walking Machine ◽  
Planning Scheme ◽  
Increasing Demand

Abstract The manufacture and maintenance of large parts in ships, trains, aircrafts, and so on create an increasing demand for mobile machine tools to perform in-situ operations. However, few mobile robots can accommodate the complex environment of industrial plants while performing machining tasks. This study proposes a novel six-legged walking machine tool consisting of a legged mobile robot and a portable parallel kinematic machine tool. The kinematic model of the entire system is presented, and the workspace of different components, including a leg, the body, and the head, is analyzed. A hierarchical motion planning scheme is proposed to take advantage of the large workspace of the legged mobile platform and the high precision of the parallel machine tool. The repeatability of the head motion, body motion, and walking distance is evaluated through experiments, which is 0.11, 1.0, and 3.4 mm, respectively. Finally, an application scenario is shown in which the walking machine tool steps successfully over a 250 mm-high obstacle and drills a hole in an aluminum plate. The experiments prove the rationality of the hierarchical motion planning scheme and demonstrate the extensive potential of the walking machine tool for in-situ operations on large parts.

Uncertain Parameters in Thermal Machine-Tool Models and Methods to Design their Metrological Adjustment Process

2015 ◽  
Vol 794 ◽  
pp. 379-386 ◽  
Author(s):  
Bernd Kauschinger ◽  
Steffen Schroeder
Keyword(s):  
Machine Tool ◽  
Design Process ◽  
Machine Tools ◽  
Expert Knowledge ◽  
Sufficient Accuracy ◽  
Adjustment Process ◽  
Uncertain Parameters ◽  
Systematic Analysis ◽  
Adjustment Procedure ◽  
Typical Design

The measures taken to improve the thermal behaviour of machine tools are based on thermal models. The models are applied to support the design process and to correct the machine tool operation in a control-based way. Especially the models for correction purposes include uncertain parameters that cannot be estimated with sufficient accuracy. Thus these parameters have to be adjusted by means of measurements. During the adjustment process, a broad diversity of machine behaviour and model characteristics has to be taken in to account. Therefore, substantial time, effort and expert knowledge are required. To identify the key expenses, a generalized and systematic analysis of the adjustment process was carried out. First, the typical design of the models, the parameters of the sub models and the current adjustment procedure were investigated. Based on the results of the analysis, support requirements were identified. Afterwards first methods and software tools for efficient support were developed. This strategy is demonstrated using the example of a hexapod strut model.

Sign in / Sign up

Export Citation Format

Share Document