Now that collaborative robots are becoming more widespread in industry, the question arises how we can make them better co-workers and team members. Team members cooperate and collaborate to attain common goals. Consequently they provide and receive information, often non-linguistic, necessary to accomplish the work at hand and coordinate their activities. The cooperative behaviour needed to function as a team also entails that team members have to develop a certain level of trust towards each other. In this paper we argue that for cobots to become trusted, successful co-workers in an industrial setting we need to develop design principles for cobot behaviour to provide legible, that is understandable, information and to generate trust. Furthermore, we are of the opinion that modelling such non-verbal cobot behaviour after animal co-workers may provide useful opportunities, even though additional communication may be needed for optimal collaboration. Marijke Bergman, Elsbeth de Joode, +1 author Janienke Sturm Published in CHIRA 2019 Computer Science
MULTIFILE
This paper describes some explorations on the concept of disassemblability as an important circularity indicator for products because of its severe impact on reuse value. Although usefulness of the concept for determining disassembly strategies and for improving circular product design clearly shows in earlier studies, the link with Industry 4.0 (I4.0)-related process innovation is still underexposed. For further technical development of the field of remanufacturing, research is needed on tools & training for operators, diagnostics, disassembly/repair instructions and forms of operator support. This includes the use of IoT and cobots in remanufacturing lines for automatic disassembly, sorting and recognition methods; providing guidance for operators and reduction of change-over times. A prototype for a disassembly work cell for a mobile phone has been developed together with researchers and students. This includes the removal of screws by means of a cobot using both vision & the available info in the product’s Bill-Of-Materials, the removal of covers, opening of snap fits and replacement of modules. This prototyping demonstrates that it is relatively easy to automate disassembly operations for an undamaged product, that has been designed with repairability in mind and for which product data and models are available. Process innovations like robotisation influence the disassemblability in a positive way, but current indicators like a Disassembly Index (DI) can’t reflect this properly. This study therefore concludes with suggestions for an evaluation of disassemblability by looking at the interaction between product, process and resources in a coherent way.
MULTIFILE
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
Collaborative robot arms (cobots) are gaining a strong foothold in contemporary manufacturing workplaces. While more information about the cobot’s impact becomes available, crucial design, work perception, performance, and strategic implications are systematically overlooked. Following a modern sociotechnical systems design theory (MSTS) perspective, which lies at the heart of workplace innovation literature, we studied if, how, and why the cobot made production units more resilient and strategically relevant. We ran a comparative case study involving 15 Dutch small- and medium-sized manufacturing enterprises (SMEs) and 36 interviewees (managers and operators). The results describe how the cobots are designed as autonomous and rigid mini-robots, handling one or a few high-quantity products in ways that are not inherently more reliable and efficient. Operators interacting with the cobots experience stronger motivational work characteristics, but the cobot’s autonomous and stable operation also provokes classic out-of-the-loop problems. Consequently, cobot-equipped production units do not always perform better. Nonetheless, SMEs deem their units strategically relevant since they (indirectly) improve financial flexibility, increase production capacity, streamline future automation projects, and accommodate the resolution of labor scarcity issues. This research creates a pathway for more MSTS and workplace innovation research at the crossroads of human-robot interaction, organisational design, production management, applied psychology, and entrepreneurship. Practical implications are provided and discussed elaborately.
MULTIFILE
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
It is much easier to run our projects right (project management) than to do the right things, or ask the right questions (leadership). also for me and my research group. We proudly report that cobots are gaining ground and that this is understandable given the major staff shortages and the need for companies to operate more efficiently and effectively. In this article it is argued that in fact,it may be understandable, but it is just not the good thing to do!
MULTIFILE
Dutch industrial manufacturers are confronted a new and promising industrial robot: the collaborative robot (cobot). These small robotic arms are revolutionary as they allow direct and safe interaction with production workers for the very first time. The direct interaction between production worker and cobot has the potential to not only increase efficiency, but also enhance flexibility as it can align the strengths of (wo)man and machine more thoroughly. Currently, Dutch manufacturers are experimenting with cobots. To obtain a first understanding about the use of cobots in Dutch industrial practice and what the consequences are for operators and production work, we conducted an exploratory interview study (N=61). We learnt that most cobots under study are used for the production of one or a few large product batches (mass production) and work highly autonomous. The interaction between cobot and production worker is limited and reduced to operators preventing the cobot from falling into a standstill. The results tend to be in line with traditional industrial automation practices: an overemphasis on leveraging the technology’s potential and limited attention for the production workers’ work design and decision latitude. HR professionals were not involved and, therefore, miss out on a crucial opportunity to be of an added value.
MULTIFILE
Future work processes are going to change in several aspects. The working population (at least in Western European countries) is decreasing, while average age of employees increases. Their productivity is key to continuity in sectors like healthcare and manufacturing. Health and safety monitoring, combined with prevention measures must contribute to longer, more healthy and more productive working careers. The ‘tech-optimist’ approach to increase productivity is by means of automation and robotization, supported by IT, AI and heavy capital investments. Unfortunately, that kind of automation has not yet fulfilled its full promise as productivity enhancer as the pace of automation is significantly slower than anticipated and what productivity is gained -for instance in smart industry and healthcare- is considered to be ‘zero-sum’ as flexibility is equally lost (Armstrong et al., 2023). Simply ‘automating’ tasks too often leads to ‘brittle technology’ that is useless in unforeseen operational conditions or a changing reality. As such, it is unlikely to unlock high added-value. In healthcare industry we see “hardly any focus on research into innovations that save time to treat more patients.” (Gupta Strategists, 2021). Timesaving, more than classic productivity, should be the leading argument in rethinking the possibilities of human-technology collaboration, as it allows us to reallocate our human resources towards ‘care’, ’craft’ and ’creativity’.
DOCUMENT
This chapter explains in brief what is needed to achieve more sustainable manufacturing processes. It develops both aspects of sustainable, economic, and technical feasibility with most focus on the latter. Remanufacturing processes are described together with relevant factors that enhance their effectivity and efficiency. An overview is given of what kind of shopfloor innovations are required in the near future and some suggestions on how digital and other Industry 4.0 technologies could help to move toward circular manufacturing.
MULTIFILE
