Joaquín Ordieres-Meré

TAO-robust backpropagation learning algorithm

In several fields, as industrial modelling, multilayer feedforward neural networks are often used as universal function approximations. These supervised neural networks are commonly trained by a traditional backpropagation learning format, which minimises the mean squared error (mse) of the training data. However, in the presence of corrupted data (outliers) this training scheme may produce wrong models. We combine the benefits of the non-linear regression model tau-estimates [introduced by Tabatabai, M. A. Argyros, I. K. Robust Estimation and testing for general nonlinear regression models. Applied Mathematics and Computation. 58 (1993) 85-101] with the backpropagation algorithm to produce the TAO-robust learning algorithm, in order to deal with the problems of modelling with outliers. The cost function of this approach has a bounded influence function given by the weighted average of two psi functions, one corresponding to a very robust estimate and the other to a highly efficient estimate. The advantages of the proposed algorithm are studied with an example.

The new challenges of end-of-life tyres management systems: A Spanish case study

Directive 2008/98/EC released by the European Union represents a significant step forward in all relevant aspects of waste management. Under the already established, extended produced responsibility (EPR) principle, new policies have been enunciated to continuously achieve better overall environmental performance of key products throughout their life phases. This paper discusses how the directive is being articulated in Spain by the main integrated management system (IMS) for end-of-life (EOL) tyres since its creation in 2006. Focusing on the IMS technological, economic and legal aspects, the study provides a global perspective and evaluation of how the IMS is facing the current issues to resolve, the new challenges that have appeared and the management vision for the coming years.

Improving Manufacturing Performance by Standardization of Interprocess Communication

A number of environmental forces such as increasing value chain network complexity, decreasing product life-cycle cost, and time-to-market requirements or increasing product complexity act upon manufacturing organizations, enhancing the acute need for organizational routines that foster efficient and effective communication between processes. Such organizational routines erode quickly in the absence of common standards for knowledge sharing, that is why successful manufacturing systems benefit from interprocess standardization. The purpose of this paper is to offer a standardization model of interprocess communication that increases manufacturing operational performance (MOP). First, we propose a novel holistic model that makes standardized interprocess communication possible in manufacturing organizations. Second, we propose a model for quantifying the implications of standardizing interprocess communication upon MOP. Finally, as a matter of application, we show the results of its successful implementation in one Japanese manufacturing organization.

Development of neural network-based models to predict mechanical properties of hot dip galvanised steel coils

In the industrial arena, artificial neural networks are among the most significant techniques in system modelling because of their efficiency and simplicity. In this paper, we present an application of artificial neural networks, along with other techniques stemming from data mining, to model the yield strength, tensile strength, elongation, strain hardening coefficient and the Lankford’s anisotropy coefficient of galvanised steel coils, according to the manufacturing process data. In particular, we propose the use of these models to improve the current control systems of hot-dip galvanising lines since an open loop control strategy must be adopted because the mechanical properties of hot-dip galvanising coils are not directly measurable.

Continuous Monitoring of Essential Tremor Using a Portable System Based on Smartwatch

Introduction Essential tremor (ET) shows amplitude fluctuations throughout the day, presenting challenges in both clinical and treatment monitoring. Tremor severity is currently evaluated by validated rating scales, which only provide a timely and subjective assessment during a clinical visit. Motor sensors have shown favorable performances in quantifying tremor objectively. Methods A new highly portable system was used to monitor tremor continuously during daily lives. It consists of a smartwatch with a triaxial accelerometer, a smartphone, and a remote server. An experiment was conducted involving eight ET patients. The average effective data collection time per patient was 26 (±6.05) hours. Fahn–Tolosa–Marin Tremor Rating Scale (FTMTRS) was adopted as the gold standard to classify tremor and to validate the performance of the system. Quantitative analysis of tremor severity on different time scales is validated. Results Significant correlations were observed between neurologist’s FTMTRS and patient’s FTMTRS auto-assessment scores (r = 0.84; p = 0.009), between the device quantitative measures and the scores from the standardized assessments of neurologists (r = 0.80; p = 0.005) and patient’s auto-evaluation (r = 0.97; p = 0.032), and between patient’s FTMTRS auto-assessment scores day-to-day (r = 0.87; p < 0.001). A graphical representation of four patients with different degrees of tremor was presented, and a representative system is proposed to summarize the tremor scoring at different time scales. Conclusion This study demonstrates the feasibility of prolonged and continuous monitoring of tremor severity during daily activities by a highly portable non-restrictive system, a useful tool to analyze efficacy and effectiveness of treatment.

Competence Assessment Framework for Project Management Learners and Practitioners

The goal of this paper is to present a competence-based learning and assessment framework that can facilitate the learning of both technical and human project management skills. The proposed competence assessment method is based on the participant’s performance and value creation. The information is obtained from assessment forms that relied on at least one piece of evidence and audits conducted automatically. This method can be implemented in an internet-based simulation environment. As the framework that was constructed was tested by students from various universities and locations, this work presents the results obtained, the relationships observed among four transversal competences -teamwork, leadership, motivation and results orientation– and the relationships between those competences and other variables such as the project’s final success, the effort ratio and the size of the project team.

Detection and analysis of Tremor using a system based on smart device and NoSQL database

Tremor is the most common symptoms of Parkinsons Disease (PD) and Essential Tremor (ET). Its detection and analysis during daily living plays a crucial role in the treatment of PD and ET patients. It is typically assessed in the clinic with certain tremor rating scales, which are qualitative, subject-dependent and do not necessarily reflect the real situation of the patient. In this paper, a system composed of a smartwatch, a smartphone and a NoSQL database sever is used to monitor the movements of the patients. A novel data analysis method is proposed to detect tremor and identify the connected actions. Tremor can be detected on the basis of the movement frequency difference and voluntary actions can also be recognized based on the rich information from the collected data. It helps clinicians to analyze the relationship between the tremor and a certain action. A series of simulated experiments are conducted to demonstrate the feasibility of the proposed system and data analysis method. The result shows that tremor happened during different situations can be detected with an adequate accuracy with the data collected by the proposed system. The actions around the tremor can also be identified.

An ICT based project management learning framework

The objective of this paper is to present a framework that can facilitate the university level learning process in the Project Management of different students who are enrolled in different universities in different locations and attending their own Project Management courses, but running a virtual experience in executing and managing projects. The framework includes both information systems and methodological procedures that are integrated in the information system, making it possible to assess learning performance.

An Online Assessment and Feedback Approach in Project Management Learning.

This work presents an online system to facilitate the assessment and feedback in project management education. Students are involved in real-world engineering projects in order to promote professional project management learning. Thus, students share an experience in executing and managing projects and are able to put into practice different skills and competences that a project member should possess in the development of a project. The proposed system considers competence assessment through different pieces of evidence that are pertinent to each assessed competence. Information from the three main actors in learning activities (teacher, peer, and learner) is collected by means of specifically developed online forms. All the gathered evidences are considered in a weighted integration to yield a numerical assessment score of each competence that is developed for each student. Furthermore, three different types of feedback are implemented and provided several times in order to promote and improve students’ learning. Data analysis from a specific academic course suggest that the presented system has a positive impact on students’ academic performance.

Lean Structural Network Resilience

Lean Leaders may face a tremendous amount of resistance when implementing and sustaining Lean in their organizations. The complexity associated with the dynamic of organizational processes in the 21st century: mobility of the work force, ever-changing product portfolios, and their related value stream adjustments are some of the reasons for this. Taking an organizational network view, this paper provides leaders with both a definition of Resilience, as well as coherent criteria to quantify the Lean Structural Network Resilience (LSNR) to the lean transformation that is associated with the mentioned changes in the organization. By implementing LSNR as a metric to measure resistance to change to Lean, Leaders can make informed Value Stream related business decisions in order to support sustainable growth.