Search

The paper describes the design and implementation of an open source machine vision platform for visual robot guidance and automated product inspection in manufacturing, based on OpenCV library. Using this platform, the rigid, industry-specific organization of material flows can be relaxed, shop floor resources becoming reality-aware. The main functionalities are: acquisition of video streams from multiple sources, image analysis at scene level, part recognition, locating and interaction with industrial equipment using standard, open communication protocols. The paper describes design aspects: system architecture, data acquisition and standardization of the image representation used by the analysis algorithms and object recognition module and input/output interaction protocols for a set of predefined cases. Results are reported for an implementation of the platform using a commercial image acquisition device and an industrial robot.

IF=3.954; International audience; This editorial introduces the special issue in the Elsevier journal Computers in Industry that analyses how the digital transformation of manufacturing is speeded up by two important drivers: cloud services and resource virtualization, which are vital for implementing the main building blocks - Cyber Physical Production Systems and Industrial Internet of Things - in the “Industry of the future” framework. The context of this special issue is firstly presented, with a specific focus on the federative concept of Industry 4.0. A framework characterizing research activities led in the field of the digital transformation of manufacturing processes and systems is then introduced. This framework is used to present and position the 12 papers composing the special issue. Perspectives are finally introduced as a guideline for future work in the digital transformation of manufacturing through cloud services and resource virtualization

Digital control of product-on-pallet distribution is organized on two layers in logistics systems: the low-level local process automation and the supervision and coordination dual logistic execution system. The paper's scientific contribution consists in a holonic control system for multiple robot palletizing cells for product-on-pallet distribution. In this holonic control approach, data-driven digital twins (DT) are created to optimize palletizing schedules, controlled with situation awareness and resource health monitoring. A model-driven DT configuration with embedded simulation faster than real-time is combined with the data-driven DT in a meta-level control scheme offering predictive situation awareness. The paper also presents a pure model-driven DT mock-up of a robot palletizing cell - simulation with software in the loop - in which the actual holonic control system including digital twins and decision making is used in the simulation as it will be deployed in reality. This configuration is used to design the robot palletizing cells, to experiment in limit conditions, to fine tune and validate the robot control solution.

This volume gathers the peer-reviewed papers presented at the tenth edition of the international workshop on Service Oriented, Holonic and Multi-Agent Manufacturing Systems for Industry of the Future—SOHOMA’20 organized on 1–2 October 2020 by Arts et Métiers ParisTech in collaboration with University Politehnica of Bucharest (the CIMR Research Centre in Computer Integrated Manufacturing and Robotics), Université Polytechnique Hauts-de-France (the LAMIH Laboratory of Industrial and Human Automation Control, Mechanical Engineering and Computer Science) and Polytechnic Institute of Bragança (the CeDRI Research Centre in Digitalization and Intelligent Robotics). The main objective of SOHOMA workshops is to foster innovation in smart and sustainable manufacturing and logistics systems by promoting concepts, methods and solutions addressing trends in service orientation of agent-based control technologies with distributed intelligence. The book is structured in eight parts that correspond to the technical sessions of the workshop’s program and include papers describing results of the research addressing the development and application of key enabling technologies (KET: production-, digital- and cyber-physical technologies) for the industry of the future. In concurrence with this vision of future manufacturing, the eight sections of the book address control and organization problems in the manufacturing value chain and offer smart solutions for smart factories networked in the cloud, implemented in cyber-physical systems with all resources integrated, sharing information and infrastructures, collaborating, adapting to reality and self-configuring at runtime for efficiency, agility and safety. These subjects are treated in the book’s Part 1: Cloud Networked Models of Knowledge-based Intelligent Control; Part 2: Digital Twins in Manufacturing and Beyond; Part 3: Holonic and Multi-Agent Process Control; Part 4: Ethics and Social Automation in Industry 4.0; Part 5: New Organizations based on Human Factors Integration in Industry 4.0; Part 6: Intelligent Products and Smart Processes; Part 7: Physical Internet and Logistics; Part 8: Optimal Production and Supply Chain Planning. info:eu-repo/semantics/publishedVersion

The paper describes the design and implementing solution for the digital twin of a robotized 3D vision-based bin picking process. The authors describe the design of the gripper and the robot-vision control for an efficient and safe bin picking process, which is validated by simulation. The shape, dimension, and orientation of the end effector are computed to allow the robot to handle parts safely, avoiding collisions with both the bin walls and neighbouring parts. A strategy is also proposed to handle parts in layered mode.

The paper describes the design, implementation, testing and validation of an open-source machine vision framework based on OpenCV (Open Source Computer Vision) library. This framework was developed for smart manufacturing control. Material conditioning and handling processes involving industrial robots are the processes that benefit from the proposed solution. The solution offers the following functionalities: acquisition of video streams from multiple sources, image analysis, object recognition, localization and interaction with industrial equipment using standard, open communication protocols. The paper covers several design aspects: system architecture, data acquisition and standardization of the image representation to be used by the analysis algorithms and object recognition module, input/output interaction protocols, camera-robot calibration. Results are reported for an implementation of the framework using a commercial image acquisition device and an industrial robot.

This paper describes a 10-year scientific journey in the area of Cloud-based manufacturing in the SOHOMA research community. The tour started in Paris on June 20, 2011 at Ecole Nationale Superieure d’Arts et Metiers, Paris and returns here on 1st October 2020 after annual stops in Bucharest, Valenciennes, Nancy, Cambridge, Lisbon, Nantes, Bergamo and Valencia. Several stages in the evolution of Cloud manufacturing research are recalled in their historical order: vertical enterprise integration and networking; resource and product virtualization and cloud infrastructure design; batch optimization with cloud services; real time big shop floor data streaming, machine learning in the cloud for predictive production control, resource health monitoring and predictive maintenance. Major contributions of SOHOMA authors are evoked: extending the cloud computing model to on demand shop floor resource sharing, infrastructure sharing in cloud networked enterprises, MES workload virtualization, deploying cloud services in real time with virtual machine and containers, high availability solutions and software defined networking, machine learning for predictive manufacturing.

The last years showed a trend of integrating smart technologies in agriculture, starting from irrigation control based on weather forecast, and ending with automated greenhouse control for the entire plant lifecycle using robots. The paper presents a solution for quality control and monitoring vegetables in greenhouses. During the lifecycle of a plant in automated greenhouses, the control and monitoring of the environment (soil humidity, temperature, ventilation, etc.) is not sufficient; the plants’ condition is very important and in most cases it can give more valuable feedback than the environment. This paper presents a solution for monitoring the health state of tomato plants in greenhouses, which allows detecting diseases in order to prevent their spread and also the removing or isolating affected tomatoes during the harvest. The paper gives information about the technologies which have been used and the system architecture; experimental results are reported and potential extensions of the proposed solution are described.

The paper presents a model for smart control of large scale manufacturing systems, in which pools of shop floor and computing resources are shared in a dual cloud pattern. The proposed architecture uses the holonic manufacturing paradigm by decoupling the decision layer from the control one. The decision layer uses intelligent agents that reconfigure optimally in real time the resource allocation and scheduling of operations on products at batch level; also, the resource health is monitored continuously. Decisions are taken in the high layer of the MES based on real time machine learning algorithms that predict resource performances and QoS influencing usage costs, classify and cluster resource states to predict anomalies in behaviours and prevent resource failures. The distributed control layer keeps reality awareness during production by using digital twins replicated for all resources. Data is collected in real time streams from physical resource and process twins, aggregated in time series and sent to the intelligent agents in the cloud without delaying production. Experiments discuss the forecast of abnormal pick-and-place robot operations.

The paper describes a semi-heterarchical manufacturing control solution based on a private cloud infrastructure which collects data in real-time from intelligent devices associated to shop-floor entities. The entities consist of industrial resources and mobile devices embedding the work in process on products during their manufacturing cycle. The proposed control system is developed using a common database in the cloud which handles operation synchronization and production control logic. The database component tables and the update processes are described in the article. The main functionalities of the control system are: manufacturing system configuration, control, monitoring, and optimization, and storage of historic data. An implementation framework and experimental results for the evaluation of consumed energy are reported.

The paper extends the hybrid control for the continuous radiopharmaceuticals production processes to supervisory control in order to add new functionalities and behavioural modes: optimal offline planning of the production process, adjusting on line the parameters of the process in current execution, and conditioning the process execution by the evolution of the environment context. The present research demonstrates the possibility of applying directly the holonic control paradigm to hybrid supervised control of continuous processes, exemplified by the production of radiopharmaceuticals. The design of an intelligent control solution for this type of processes uses the holonic hybrid supervised control architecture with optimally planned demand sequence which groups demands into commands and orders which are derived from a master product recipe. Implementing solutions are provided.

This paper discusses the main security-related challenges raised in distributed Manufacturing Execution System (MES) architectures. In contrast to monolithic control architectures, where security i...

In the context of today’s interconnected world, public higher education is confronted with a shift of its role from public good to public service, which is accompanied by new challenges in building and maintaining the viability of public universities. There is a vast literature on this topic, but there is not yet reported a systematic approach in modeling the basic steps of decision making and problem solving to be completed by a public university to remain viable. Starting from the concepts of service system, and consonance and resonance as key viability conditions, this paper proposes a discrete event model of the viability building and maintaining process in a generic public university organization. The model evolution is driven by events conditioned by internal, external, or mixed causes. Some of the uncontrollable events occurring in the process may drive the discrete evolution to disturbance rejection states, where specific recovery strategies are applied. If recovery is successful, disturbances may become sources of innovation. The discrete event model can be used as a theoretical and conceptual tool for building and testing instances of the viability building and maintaining processes in public universities, and also for refining academic evaluation criteria by the authorities.

The paper describes the architecture design and implementing solution for the digital twins of industrial robot, aggregated and embedded in the global health monitoring, maintenance and control system of manufacturing resources. Manufacturing scheduling and control system. The main functionalities of the digital twin are: monitoring the current status and quality of services performed by robots working in the shop floor, early detecting anomalies and unexpected events to prevent robot breakdowns and production stops, and forecasting robot performances and energy consumption. Machine learning techniques are applied in the cloud layer of the virtual twin for predictive, customized maintenance and optimized robot allocation in production tasks. The paper introduces a framework integrating the virtual robot twins in an ARTI-type control architecture, proposes a solution to implement the twin on a distributed cloud platform and exemplifies the concepts in a shop floor case study with SCARA assembly robots.

Starting from the generic, activity-oriented Service System lifecycle model, the paper considers the service as a flow of connected activities having a discrete event nature and being formalized as discrete event systems. The service activity flow is optimized off-line by a discrete supervisor and monitored at delivery by a logical controller in a 4-layer embedded digital twin architecture. The supervisor is developed using an ARTI-type holonic architecture in which operant resources (skills, knowledge) are considered intelligent agents acting on operand resources virtualized by their associated digital twins. Because they offer reality-awareness, these virtualized entities form the “intelligent beings” layer of a holonic supervisor with distributed intelligence. The holonic approach offers new perspectives for the service management: optimization and reality awareness.

Companies try nowadays to get the most technological advantages in order to be successful on the market. One of the tools they use is cloud computing; many companies are not only using a cloud infrastructure or service but more, it is estimated that until 2021 more than 90% of the companies will use multi-cloud systems and services (that means more than 5 cloud systems or services). This trend is affecting also the manufacturing companies; the challenge in this context is to manage these complex systems and services. The cloud systems and services are public, private or hybrid being offered by different providers: IBM, RedHat, Azure, Amazon Web Services (AWS) and so on. The paper presents a comparison for two cloud solutions which offer services for robotics and manufacturing. The paper also discusses the term cloud manufacturing and cloud robotics which many times is wrongly interpreted. Finally a solution for cloud robotics/manufacturing is proposed and tested for two variants of cloud solutions: using containers and virtual machines.

The paper describes the architecture design and implementing solution for the digital twin of a shop floor transportation system embedded in the global manufacturing scheduling and control system. The products are assembled on pallets travelling on the conveyor between workstations, where assigned resources perform scheduled operations. The main functionalities of the digital twin are: mirroring the current stage of the physical pallet transportation process and the state of the physical conveyor components, predicting the values of the pallet’s transportation times along the conveyor’s segments between any two workstations, applying these values for enhanced reality-awareness of optimized product scheduling and resource allocation, and detecting anomalies in the behaviour of the conveyor equipment. Starting from a shortlist of generic scenarios, AI techniques are applied in the cloud layer of the virtual twin to optimally schedule products and early detect conveyor anomalies in the context of predictive maintenance.

The paper defines the structure and utilisation mode of an embedded aggregate digital twin for the hybrid supervised control of semi-continuous production process. Multiple twin aggregations are defined for the production of radiopharmaceuticals: (i) virtual twins of three main sub processes in the production plant; (ii) predictive and decision making twins and their projections in the supervisor for process resilience and global optimization. The control system is organized according to the holonic reference architecture ARTI that enables the concepts of embedded and networked digital twins that provide collective and predictive situation awareness thus bringing closer the software control to the real process evolution. The research reported in this paper demonstrates that the holonic paradigm, first applied to discrete manufacturing process control, can be adapted to the hybrid supervised planning and control of semi-continuous production processes and brings the benefits that holonic organisations provide to living organisms: direct connectivity between physical plant entities (resources, activities and outcomes) and their virtual twins, global optimization, robustness, process resilience and plant sustainability.

International audience; This chapter gathers a number of reviews that offer an open, scientific discussion on the transition from PROSA, the well-known reference architecture for Holonic Manufacturing Systems proposed twenty years ago by the School of Leuven, Belgium, to ARTI – the new reference model for IT-based general-purpose control systems proposed by Paul Valckenaers, that stimulates a higher level of abstraction and genericity in understanding and inferring reality. ARTI models architecture elements of reality as intelligent beings in a dominant position relative to their related decision-making components modelled as intelligent agents. The reviews start with overviews of the creation process, the structural elements - holon classes and the features - aggregation, specialization, abstraction levels, flexibility and scalability. The needs that led to the evolution of PROSA into ARTI are analysed: the difficulties to spread the architecture in the research community and industry, the lack of suitability to act as reference architecture for Cyber-Physical Production Systems and Industrial Internet of Things framework in the Industry 4.0 vision of future manufacturing. The reviews discuss the elements and coverage of ARTI, and the translation of basic and staff PROSA holons into new abstract categories: intelligent beings-agents, types-instances, and resources-activities. It is appreciated that, beyond the problem of terminology of PROSA holons difficult to apply in other domains than manufacturing, the move to ARTI assures in-depth interoperability, access to the world-of-interest through digital twins that interact with smart control systems influencing thus these intelligent beings (e.g., plant, shop floor, healthcare system, environment facility, supply chain, service system) into the desired behaviours. From the software implementing point of view, most of the reviewers agreed on the fact that none of the programming languages except for Erlang provided acceptable levels of viability and sustainability for PROSA application classes. An interesting point of view is formulated about the possibility to use ARTI for Enterprise Integration and Interoperability. The present discussions appreciates the new ARTI reference architecture together with D-MAS pattern as an enhanced model to build flexible, intelligent and robust control systems based on prioritizing situation awareness and real-world interoperability. This scientific discussion about the new ARTI model, initiated, presented and critically analysed for the first time in the SOHOMA scientific community needs to be continued by the research community for enrichment and industry adoption.

Editorial du Numéro spécial : Emerging ICT concepts for smart, safe and sustainable industrial systems; International audience; This editorial introduces the special issue on Emerging Information and Communication Technology (ICT) concepts for smart, safe and sustainable industrial systems in the Elsevier journal Computers in Industry. The 13 papers in this special issue were selected because of their high quality and also because they propose emerging ICT solutions that address at least one of the three dimensions we suggest are basic requirements to design usable future Industrial Systems that must be safe, smart and sustainable. Previous global discussions about the state of the art with regard to the topic of this special issue are provided, as well as exploratory guidelines for future research in this area.

We introduce the vMES virtualization layer for manufacturing execution systems.Workload identification at MES layer according to ISA-95.03 functions.We present a process for private cloud redundancy and scaling for MES workload.Experimental results showing perturbation handling in cloud environment.Benefits of virtualization and programmable infrastructure for MES workloads. Virtualization of manufacturing execution system (vMES) workloads offers a set of design and operational advantages to enterprises, the most visible being improved resource utilization and flexibility of the overall solution. This paper explores redundancy and scalability, as other important operational advantages introduced by the use of private clouds for MES virtualization in the context of the programmable infrastructure (PI) concept. PI is a new architectural approach in which the computing infrastructure, represented by resources, networks, storage, becomes dynamic and is controlled by the application, in contrast with traditional architectures where the application has to adapt to a static infrastructure. For MES applications, the adoption of PI has the potential to add a new layer of flexibility and optimization by allowing quick configuration and re-configuration based on environmental changes, especially in the context of virtualization in private cloud where workloads can be provisioned and de-provisioned in real time. In this context, this paper presents the main redundancy and scalability requirements for the workloads identified in ISA-95.03 based solutions and discusses in detail the strategies to assure the redundancy and scalability requirements of these workloads both individually and at the system level. The main contributions of this paper are therefore the introduction of PI combined with private cloud virtualization at the MES layer in order to achieve redundancy and scalability of the control solution. The pilot implementation presented is based on PI concepts and is realized in practice using SOA BPEL and IBM CloudBurst REST APIs. The MES system considered for the pilot implementation adopts a multi-agent vMES architecture having COBASA-type functionality. The experimental results presented in this paper show the system response in a set of failure scenarios, with focus on the reconfiguration time of workloads, and the dynamic response to perturbations in the system.

The digitalization processes in manufacturing enterprises and the integration of increasingly smart shop floor devices and software control systems caused an explosion in the data points available in Manufacturing Execution Systems. The degree in which enterprises can capture value from big data processing and extract useful insights represents a differentiating factor in developing controls that optimize production and protect resources. Machine learning and Big Data technologies have gained increased traction being adopted in some critical areas of planning and control. Cloud manufacturing allows using these technologies in real time, lowering the cost of implementing and deployment. In this context, the paper offers a machine learning approach for reality awareness and optimization in cloud. Specifically, the paper focuses on predictive production planning (operation scheduling, resource allocation) and predictive maintenance. The main contribution of this research consists in developing a hybrid control solution that uses Big Data techniques and machine learning algorithms to process in real time information streams in large scale manufacturing systems, focusing on energy consumptions that are aggregated at various layers. The control architecture is distributed at the edge of the shop floor for data collecting and format transformation, and then centralized at the cloud computing platform for data aggregation, machine learning and intelligent decisions. The information is aggregated in logical streams and consolidated based on relevant metadata; a neural network is trained and used to determine possible anomalies or variations relative to the normal patterns of energy consumption at each layer. This novel approach allows for accurate forecasting of energy consumption patterns during production by using Long Short-term Memory neural networks and deep learning in real time to re-assign resources (for batch cost optimization) and detect anomalies (for robustness) based on predicted energy data.

The paper applies the holonic paradigm to the supervised hybrid control of semi-continuous processes, which are exemplified by the production of radiopharmaceutical substances. The supervisor of the control system fulfils two main functionalities: (i) optimization of global process planning that includes all client orders most recently received (the values of process parameters and operations timing are initially computed to maximize the number of accepted orders—optimal state trajectory); (ii) reconfiguring the parameters of the optimal state trajectory whenever unexpected events occur (in the production sub processes or in the environment parameters), providing thus robustness at disturbances. The implementation of the holonic supervised hybrid control system uses the multi-agent framework in semi-heterarchical topology. Two scenarios validating the optimization of planning and experimental results are reported.

The paper presents a solution for data collection from devices embedded on industrial resources. The proposed architecture is validated using an industrial robot from Omron. Since not all devices that acquire electrical signals from sensors and transmit them in centralized environments (e.g., Cloud) have advanced processing capabilities, the creation of an aggregation node that concentrates data from different types of sources and sends it to the Cloud database is proposed. The data is collected, aggregated and uploaded to the private cloud platform for centralized manufacturing control tasks (product scheduling, resource allocation, monitoring and diagnosis). Experimental results are reported.

Large scale manufacturing systems are capable to execute manufacturing operations across multiple product batches by coordinating many shop floor actors. Monitoring and processing in real time the information flow from these systems becomes an essential part in optimizing and detecting faults that might affect the production schedule. This paper proposes an architecture that uses big data concepts and map-reduce algorithms to process the information streams in large scale manufacturing systems, focusing on energy consumptions aggregated at various layers. Once the information is aggregated in logical streams and consolidated based on relevant metadata, a neural network is trained and used to learn historical patterns in data on each layer. This novel approach also allows accurate forecasting of the energy consumption patterns during the production cycle by using Long Short Term Memory neural networks. The paper presents a practical example on how map-reduce algorithms can be implemented and how repetitive patterns in energy consumption can be learned.

Edge computing is essential for the Industrial IoT as framework for data acquisition from shop floor devices; distributed intelligence will shift to the edge for speed reasons in real-time handling of big data. This research aims at developing a generic architecture for information and data collection, smart processing and aggregation at the edge of large-scale manufacturing control systems; the edge is represented by the set of shop floor entities (things) – resources and intelligent products that are agentified and communicate in multi-agent systems for decentralized MES tasks. The IIoT architecture integrates a private cloud platform with a network of IoT aggregation nodes composed of IoT gateways, sensors and PC-type workstations hosting the resource agents. Both networks form the distributed MES layer of a semi-heterarchical, cloud-based production control system. The implementing solution is given; experiments report communication with the cloud.

Manufacturing control systems evolved in the recent decades from pre-programmed rigid systems to adaptable, data driven, cloud based implementations, capable to respond to environment changes and new requirements in real time. A byproduct of this transformation is represented by large amounts of structured and semi-structured information, both historical and real-time data that is made available on various layers of the system. This accumulation of information brings the opportunity to move from the rule based decision making algorithms used traditionally by these control systems towards more intelligent approaches, driven by modern deep learning mechanisms. This paper proposes a time series forecasting model using recursive neural networks (RNN) for operation scheduling and sequencing in a virtual shop floor environment. The time series aspect of the RNN is novel in manufacturing domain, in the sense that the new best prediction produced considers the previous decisions and outcomes. The proposed implementation explains how the RNN can be mapped to the specifics of a manufacturing control system and introduces a bidding mechanism to allow dynamic evaluation of individual forecasts. The pilot implementation, initial experiments on sample data sets and results presented show how using recursive neural networks can optimize resource utilization and energy consumption.

This paper shows how service science principles may be used for engineering and realizing improved public transport services. It approaches a value co-creation perspective for the management of public transport service operations based on an activity-based model of a generic service system that allows capturing requirements for software intensive service systems. The main focus is on specific implementation issues of the activity-based model of the generic service system, with a strong accent on its most representative component, the service set-up and configuring unit. This model is applied in a case study for planning of a public transport service and describes how a specific service reconfiguring request is formulated. This examination is further used as a document of requirements that drives the construction of an agent-based model expressing value-creation interactions among service system’s stakeholders in public transport services. The usefulness of the developed agent-based model for the analysis of service systems operational capabilities is suggested through simulation. A business scenario related to the management of public transport services is described, and the defined agent-based model is executed with the Presage2 multi-agent programming platform in order to capture specific issues of piece-of-work planning. The proposed approach, evaluated on the simple working scenario, fosters the role of service interaction modeling in supporting a public transport service system to dynamically adapt its operational capabilities in delivering good public transport services, as more or less quantifiable changes can affect service delivery over time.

The paper proposes an agent-based approach for measuring in real time energy consumption of resources in job-shop manufacturing processes. Data from industrial robots is collected, analysed and assigned to operation types, and then integrated in an optimization engine in order to estimate how alternating between makespan and energy consumption as objective functions affects the performances of the whole system. This study focuses on the optimization of energy consumption in manufacturing processes through operation scheduling on available resources. The decision making algorithm relies on a decentralized system collecting data about resources implementing thus an intelligent manufacturing control system; the optimization problem is implemented using IBM ILOG OPL.

Capacity measurement is of crucial importance in business and manufacturing and is intimately related to both finite and infinite capacity planning. There is a vast literature on this subject and capacity may be defined in various ways. This paper investigates, within two small scale examples, some issues regarding capacity measurement in a shop floor with multifunctional parallel machines that have to process a specified quantity of products of different types, and with specific operations requirements, in a specified working time. In the first example, each operation type has a specific operation unit time, independently on the working machine, while in the second example each machine can execute any operation from its capabilities portfolio in a unique operation’s unit time, with a unique associated operation unit cost. The study emphasizes, in the first example, that the capacity measurement depends not only on machines capabilities, on products requirements and on the imposed working time, but also on the allocation strategy of groups of machines to groups of products, while the second example shows that, in case of machines with unique operation unit time for all operation types, the maximal number of operations executable in given working time is a valid capacity measure. The discussed examples may serve as starting point for defining capacity planning procedures for more complex scenarios that can be tested using dedicated software tools, targeting industrial applications.

Real time analysis of data collected from the shop floor opens the path towards efficient scheduling of batch execution for large scale distributed manufacturing systems. Prediction of the shop floor activities has a great potential to reduce manufacturing costs, by providing the information required for operational decisions like preventive maintenance, automatic remediation or scheduling optimization. Research has been focusing on how machine learning algorithms can be used to better understand and extract insights from historical data collected from manufacturing systems. However, in the current manufacturing environments, driven by mass customization and short time to market, these approaches fail to be agile enough to be useful. In this paper we propose a real-time machine learning approach for large scale manufacturing systems that can predict various scenarios before service degradation occurs, thus allowing for corrective actions. At the same time, outliner detection algorithms can be used to evaluate the system’s health at a holistic level. Scalability requirements are achieved by modelling the architecture around data streams processed in real time by map-reduce operations. The concepts presented in this paper build on recent developments on flexible, distributed and cloud based manufacturing, where these real time actions can be efficiently implemented.

In many hospitals all over the world there is an acute lack of physicians; in addition, doctors who are working in hospitals are overwhelmed by the number of patients and other administrative duties which they must do. Due to the specific of the work many operations/procedures/activities must be done manually and there are no automated systems which could improve the quality of the medical service and the efficient usage of the physician’s time. In this paper we propose a service system designed to help physicians to automate the work of registering patient clinical observations into the patient clinical observation sheet. The procedure of registering observations can be time consuming in some situations due to the numerous parameters which must be registered. The proposed system uses voice to text conversion engine to register the observations; thus, doctors spend much less time to review the clinical observations and eventually make corrections if necessary.

The paper proposes a methodology for replicating in the cloud software agents associated to the control of manufacturing resources. Replicating in Cloud Manufacturing Control architectures (CMfg) agents and their services results in a high availability (HA) decentralized control system. Agents’ services and replicated data will be detailed in the paper. This methodology represents an extension of the generic agentification process which consists in associating a software agent to a physical entity in order to simplify the access to the resource’s operations managed as services and easily accessed through standard messages in multi-agent control frameworks (MAS). The developed methodology is validated using the JADE framework. The paper explains how a JADE agent acts as intermediary between the MAS framework based on the exchange of standardized FIPA messages, and direct resource communication which is based on exchanging information over a TCP connection.

The large scale emergence in the last decade of various cloud solutions, ranging from software-as-a-service (SaaS) based solutions for business process management and implementation to very sophisticated private cloud solutions capable of high performance computing (HPC) and efficient virtualization, constitute the building blocks for engineering the next generation of flexible enterprise systems that can respond with great agility to changes in their environment. These new technologies are adopted at a certain level by manufacturing enterprises in order to advance in a new era of mass customization where flexibility, scalability and agility are the differentiating factors. In this context, this paper introduces the virtualized manufacturing execution system (vMES), an intermediate layer in the manufacturing stack, and discusses the advantages and limitations offered by this approach for manufacturing enterprises. A classification of MES workloads based on the ISA-95 function model is presented, focusing on the virtualization techniques suitable for each workload, considering the algorithms and technologies used and the virtualization overhead. A pilot vMES implementation using a parallel process for smart resource provisioning and automatic scaling is also presented. The pilot implementation using six Adept robots and one IBM CloudBurst 2.1 private cloud and an ISA-95 based MES is described; the virtualization sequence is analyzed in several scenarios of resource workload collocation on physical cloud blades with and without perturbations.

Multi agent systems represent an elegant approach for the control architecture of manufacturing systems. Distributed control architectures have the potential to achieve greater flexibility by being capable of local decision making based on real time reasoning. One of the main challenges of these distributed architectures is represented by the capability to synchronize the production data across all execution points in a reliable and consistent fashion. In this context, this paper aims to resolve the problems associated with real time production data synchronization in distributed multi-agent control systems by proposing a common dataset synchronized across all agent entities using multicast network communication. On top of this common dataset approach, an agent negotiation mechanism is proposed that addresses the operation sequencing and resource allocation in decentralized operation model. The pilot implementation is using JADE multi agent platform and JGroups for real time data synchronization and NetLogo for abstract representation of the simulation system. Experimental results gathered from the pilot implementation are discussed.

Starting from a hypothetical open shop production quantity estimation problem with weakly formulated objective function, this paper comparatively discusses a manual heuristic solution and an IBM-ILOG based numerical solution. Problem complexity is managed by specific partitioning strategies in the construction of heuristic solution instances. The study may serve as starting point in the systematic building of a more general heuristic solution method with potential applications in business plan forecasting.

The paper describes a computational solution with cloud implementing, extending robot-vision capabilities of real-time multiple articulated objects recognition for on-the-fly robot grasping. Articulated objects are recognized by matching the unknown object’s skeleton computed from the input image in a cloud virtual machine (VM) with a set of learned skeleton signatures. This High Performance Computing (HPC) process represents a powerful capability for qualitative shape matching because it unambiguously synthesizes and helps estimating the topology of the object and its shape. The skeleton-based matching process is performed as an application-driven robotic service in a private cloud, ten times faster than the robot controller is able to do it and nearly twice faster than two PC-type robot terminals for multiple parts moving on conveyor belts. The parameters of the virtualization process and experimental results which confirm the solution are presented.

Large scale emergence of mature cloud solutions, ranging from software-as-a-service based solutions for business management, to very sophisticate private cloud solutions; offer the building blocks for constructing extremely flexible enterprise systems that can respond to environmental changes with great agility. Manufacturing enterprises need to adopt these new technologies to advance in a new era of mass customization where flexibility, scalability and agility are the differentiating factors. In this context, this paper introduces the virtualized MES and shop floor architecture as an intermediate layer in the manufacturing stack and discusses the advantages offered by this approach for manufacturing enterprises. A classification of MES and shop floor devices is presented focusing on the virtualization techniques suitable for each device type, considering the level of distributed intelligence and the virtualization overhead. Shop floor virtualization through shop floor profiles is presented and discussed underlying the flexibility of the solution. A pilot multi-agent implementation for virtual shop floor configuration based on the CoBASA reference architecture is presented and discussed. The shop floor profiles which define the virtual layout and mappings of the robotized manufacturing system are also provided in this context. The pilot implementation using six Adapt robots and a IBM CloudBurst 2.1 private cloud, is described and virtualization overhead in terms of event propagation delays is measured and presented in several scenarios of resource workload collocation on physical cloud blades

One of the most important problems when considering the design of manufacturing systems based on SOA paradigms is the integration of shop floor devices in the business processes at the enterprise level. This paper presents the design and implementation of the Customer Order Management (COM) module based on SOA architecture in the context of holonic manufacturing systems. The COM module is integrating with SOA enabled shop floor devices using industry standards. The implementation leverages a multi agent system suited for industrial applications integrated in a SOA environment capable of dynamic BPEL workflow generation and execution. The prototype consists in a SCA application for core COM module functionality and an extension for NetLogo MAS platform for SOA integration. The COM module interacts with the MES layer using real time events handled by the BPEL process implementation in the execution stage. A web based portal frontend for the COM module has been developed to allow real time tracking of customer orders, providing data about product batch execution and individual progress of each product on the production line.

The Intelligent Product concept has been proposed almost a decade ago, gaining considerable traction in manufacturing enterprises as it helps aligning the information flow with the material flow even in the simplest implementations. In advanced applications it can provide support for executing complex algorithms allowing local decision making upon collective reasoning or individual and autonomous selection of the manufacturing operations. In the context of service-oriented manufacturing systems there are new requirements for intelligent products, such as standards adoption and SOA capabilities. This paper proposes a classification of intelligent products from a SOA integration point of view and introduces a formalized data structure for intelligent products in the form of a XSD schema for XML representation. The data flow during manufacturing is discussed in the context of lead time and lag time between operations in the product recipe with the goal to enable ETA estimation of the product and the product batch. An example is presented for assembly of an H- shaped product, focusing on representation of operation dependencies and lag time in XML format proposed.