Industrial robot manipulators are widely used for repetitive applications that require high precision, like pick-and-place. In many cases, the movements of industrial robot manipulators are hard-coded or manually defined, and need to be adjusted if the objects being manipulated change position. To increase flexibility, an industrial robot should be able to adjust its configuration in order to grasp objects in variable/unknown positions. This can be achieved by off-the-shelf vision-based solutions, but most require prior knowledge about each object tobe manipulated. To address this issue, this work presents a ROS-based deep reinforcement learning solution to robotic grasping for a Collaborative Robot (Cobot) using a depth camera. The solution uses deep Q-learning to process the color and depth images and generate a greedy policy used to define the robot action. The Q-values are estimated using Convolutional Neural Network (CNN) based on pre-trained models for feature extraction. Experiments were carried out in a simulated environment to compare the performance of four different pre-trained CNNmodels (RexNext, MobileNet, MNASNet and DenseNet). Results showthat the best performance in our application was reached by MobileNet,with an average of 84 % accuracy after training in simulated environment.
DOCUMENT
The number of applications in which industrial robots share their working environment with people is increasing. Robots appropriate for such applications are equipped with safety systems according to ISO/TS 15066:2016 and are often referred to as collaborative robots (cobots). Due to the nature of human-robot collaboration, the working environment of cobots is subjected to unforeseeable modifications caused by people. Vision systems are often used to increase the adaptability of cobots, but they usually require knowledge of the objects to be manipulated. The application of machine learning techniques can increase the flexibility by enabling the control system of a cobot to continuously learn and adapt to unexpected changes in the working environment. In this paper we address this issue by investigating the use of Reinforcement Learning (RL) to control a cobot to perform pick-and-place tasks. We present the implementation of a control system that can adapt to changes in position and enables a cobot to grasp objects which were not part of the training. Our proposed system uses deep Q-learning to process color and depth images and generates an (Formula presented.) -greedy policy to define robot actions. The Q-values are estimated using Convolution Neural Networks (CNNs) based on pre-trained models for feature extraction. To reduce training time, we implement a simulation environment to first train the RL agent, then we apply the resulting system on a real cobot. System performance is compared when using the pre-trained CNN models ResNext, DenseNet, MobileNet, and MNASNet. Simulation and experimental results validate the proposed approach and show that our system reaches a grasping success rate of 89.9% when manipulating a never-seen object operating with the pre-trained CNN model MobileNet.
DOCUMENT
Op de hogeschool van Utrecht en de Fontys hogescholen doen twee promovendi van de Technische Universiteit Delft onderzoek naar assemblage systemen voor miniatuurcomponenten. De nadruk ligt op het assembleren van elektronica-componenten door Pick-and-Place (P&P) machines op Printed Circuit Boards (PCB's). Deze P&P machines hebben een output van enkele duizenden componenten per uur per plaatsingskop. De snelste P&P-machine in het veld (2001) is de FCM II van Assembleon met een output van 6000 componenten per uur per plaatsingskop. De plaatsings nauwkeurigheid bedraagt 100 um. Het Doel van het onderzoek is output verhoging, met minimaal een factor 2, met behoud van plaatsingsnauwkeurigheid.
DOCUMENT
Presentation given at EURCRIM 2022 conference
DOCUMENT
The principal aim of this study is to explore the relations between work domains and the work-related learning of workers. The article is intended to provide insight into the learning experiences of Dutch police officers during the course of their daily work. Interviews regarding actual learning events and subsequent changes in knowledge, skills or attitudes were conducted with police officers from different parts of the country and in different stages of their careers. Interpretative analyses grounded in the notion of intentionality and developmental relatedness revealed how and in what kinds of work domains police officers appear to learn. HOMALS analysis showed work-related learning activities to vary with different kinds of work domains. The implications for training and development involve the role of colleagues in different hierarchical positions for learning and they also concern the utility of the conceptualisation of work-related learning presented here.
DOCUMENT
Lectorale rede
DOCUMENT
Procrastination is one of the most common challenges when it comes to studying and working productively. You may not feel like studying for a while, feel demotivated, become easily distracted, feel tired, or find a certain task difficult. You might also suffer from perfectionism or fear of failure. Perhaps you are a “master procrastinator” and procrastinate with almost all your tasks. If that is the case, you probably could use some advice that will help to permanently change your study behavior. Check out these tips, which are based on scientific insights from cognitive psychology, neuropsychology, educational sciences and our own research.
DOCUMENT
We need mental and physical reference points. We need physical reference points such as signposts to show us which way to go, for example to the airport or the hospital, and we need reference points to show us where we are. Why? If you don’t know where you are, it’s quite a difficult job to find your way, thus landmarks and “lieux de memoire” play an important role in our lives.
LINK
Ethnographic fieldwork is a balancing act between distancing and immersing. Fieldworkers need to come close to meaningfully grasp the sense-making efforts of the researched. In methodological textbooks on ethnography, immersion tends to be emphasized at the expense of its counterpart. In fact, ‘distancing’ is often ignored as a central tenet of good ethnographic conduct. In this article we redirect attention away from familiarization and towards ‘defamiliarization’ by suggesting six estrangement strategies (three theoretical and three methodological) that allow the researcher to develop a more detached viewpoint from which to interpret data. We demonstrate the workings of these strategies by giving illustrations from Machteld de Jong’s field- and text-work, conducted among Moroccan-Dutch students in an institution of higher vocational education.
DOCUMENT
Ondernemen in een veranderende wereld is geschreven voor beleidsmakers, managers, ondernemers, organisatieadviseurs en studenten. Vanuit diverse ontwikkelingen op het gebied van technologische connectiviteit, open innovatie, maatschappelijk verantwoord ondernemen, outsourcing, de herrijzenis van China, samenwerking tussen organisaties, de veranderende consument, de veranderende marketing en authenticiteit, biedt Ondernemen in een veranderende wereld een nieuw perspectief op een veranderende wereld. In dit hoofdstuk wordt ingegaan op Maatschappelijk Verantwoord Ondernemen (MVO). Vaak in één adem genoemd met duurzaamheid. Geen onderneming lijkt zich meer te kunnen veroorloven er niet aan te doen. Waar komt deze ontwikkeling vandaan? Wat moeten we nu precies onder maatschappelijk of duurzaam ondernemen verstaan? En is MVO verenigbaar met meer gangbare financiële ondernemingsdoelstellingen? Vragen die een antwoord, of minstens een aanzet daartoe verdienen.
DOCUMENT
