Industry 4.0 technologies are increasingly driving the development of Zero-Defect Manufacturing (ZDM) solutions to reduce waste and enhance product quality. However, traditional ZDM methods face challenges in adaptability, real-world applicability, and handling data scarcity. This paper presents a cloud-based modular Decision-Support System (DSS) tailored for the decorative surfaced panels industry that incorporates real-time process monitoring, Machine Learning (ML), and Explainable Artificial Intelligence (XAI) to proactively predict defects, identify root causes (e.g., temperature fluctuations), and recommend optimal machine parameters. Key innovations include a Zero-Shot Learning (ZSL) component for generating initial machine recipes for new panel types, reducing setup-phase waste, and Asset Administration Shells (AASs) for interoperable data communication. Evaluations demonstrate the DSS achieves a 90% F1-score in defect classification and reduces defect rates by 20 percentage points (pp) through parameter optimization. The architecture’s modularity and integration with an industrial use case highlight its scalability and applicability across diverse manufacturing environments.

Machine learning (ML) for industrial anomaly detection often assumes access to abundant labeled data, yet realworld deployments face data scarcity due to limited observability, high labeling costs, and rare attack events. This paper presents a detailed exploration of the "M2M Using OPC-UA" dataset, a publicly available benchmark used in multiple industrial intrusion detection studies, through statistical analysis, dimensionality reduction, outlier detection, and feature importance techniques. The study analyse the use of the dataset across diverse ML paradigms, including biologically inspired, deep, and probabilistic models. Findings reveal strong class separability in the binary task but substantial imbalance in rare multi-class labels. While highly structured, the dataset underrepresents data scarcity unless multi-class formulations are considered, making it both a valuable and limited benchmark for scarcity-aware ML.

Effectively integrating knowledge-based data with data-driven techniques in hybrid digital twins (HDTs) is essential for advancing monitoring capabilities in Industry 4.0. This integration involves leveraging domain knowledge (physics) to guide model structure, inform training data, or reveal new parameters and behaviours, while data is used to refine models, improve accuracy, estimate unknown parameters, and capture phenomena not considered in theoretical approaches. To address this challenge, we propose a novel HDT framework, inspired by Landau-Ginzburg Theory (LGT) of phase transitions, that enables the digital twin to be computationally lightweight and physically interpretable. In this study, machine learning-based production data simulation was applied to a real SIF-400 smart assembly line training system, serving as a conceptual case study. This allows modelling of the operating conditions in such systems. Three operating phases — Optimal, Stressed, and Failed — are characterised by a stability function. Analysis of the corresponding stability profiles validates the framework’s ability to classify system states, highlighting its potential for holistic stability assessment and early detection of critical transitions in system dynamic processes. Bi-directional communication enables automatic system stoppage upon failure detection.

As Cyber-Physical Production Systems (CPPS) and the Industrial Internet of Things (IIoT) evolve, the integration of Artificial Intelligence (AI) and Edge Computing is enabling more responsive, autonomous, and intelligent industrial operations. However, this shift introduces a set of emerging requirements that traditional CPPS architectures are not equipped to address. This paper presents a structured requirement analysis for AI- and Edge-enabled CPPS in industrial contexts. The analysis is organized into three domains: (i) AI/ML-specific requirements, (ii) Edge Computing requirements, and (iii) Edge analytics requirements. Each requirement is characterized across dimensions such as reliability, adaptability, and resource intensity, and illustrated through realistic industrial scenarios. Grounded in insights from multiple collaborative R&D projects, the framework captures technical and operational needs, offering a practical reference for designing and deploying next-generation CPPS in smart manufacturing and beyond.

The Dendritic Cell Algorithm (DCA), inspired by the Human Immune System, is a promising Artificial Immune System (AIS) for anomaly detection. However, its pre-processing phase traditionally depends on expert manual intervention, limiting scalability and objectivity. This study investigates the impact of integrating automated feature reduction techniques to streamline this phase. We propose three approaches: Kernel Principal Component Analysis (KPCA), Autoencoders (AE), and a hybrid Autoencoder with KPCA (AEkPCA). The models were tested on seven datasets from cybersecurity, biology, and finance domains. KPCA, particularly with a Gaussian kernel, delivered the most consistent results, achieving high accuracy and strong MCAV separation. AEkPCA showed competitive performance in complex datasets, especially with polynomial kernels, though with greater variability. AE in isolation exhibited unstable behavior and inconsistent detection. These results support the viability of automated preprocessing in DCA, highlighting that performance depends heavily on the feature reduction method and kernel combination.

Esta dissertação aborda o desafio crítico de melhorar a saúde operacional e a longevidade de máquinas avançadas em sectores como a indústria, os cuidados de saúde, a defesa e os transportes.Com o nascimento da Indústria 4.0, os métodos de manutenção tradicionais provaram ser menos eficazes, levando à adoção de estratégias de manutenção preditiva.Esta mudança é crucial para melhorar a eficiência operacional e minimizar os tempos de paragem inesperados.A investigação centra-se no desenvolvimento e validação de uma ferramenta de manutenção preditiva baseada em dados, utilizando técnicas avançadas de aprendizagem automática e análise de dados para prever potenciais falhas do equipamento e otimizar os planos de manutenção.O estudo utiliza uma abordagem de um conjunto de dados duplo, utilizando dados reais de um dataset da Gorenje e de um dataset de manutenção preditiva da Microsoft Azure presente no Kaggle, para validar os modelos preditivos propostos.Os modelos avaliados incluem Cox Proportional Hazards (CoxPH), Random Survival Forests (RSF), Gradient Boosting Survival Analysis (GBSA) e Survival Support Vetor Machines (FS-SVM).Os resultados demonstram níveis variáveis de eficácia do modelo, com valores de índice C nos dados de teste de 0,792 para CoxPH, 0,601 para RSF, 0,579 para GBSA e 0,514 para FS-SVM.O FS-SVM apresentou uma melhoria significativa no desempenho do teste em comparação com os dados de treino, indicando o seu potencial para um maior refinamento.É ainda realizada uma simulação na qual é possível prever quando os componentes vão falhar assim como avaliar os custos associadas às suas reparações.Esta dissertação não só contribui para a área académica ao fornecer uma avaliação dos métodos de análise de sobrevivência no contexto da manutenção preditiva, como também introduz uma interface de painel de controlo de fácil utilização para a tomada de decisões de manutenção em tempo real.A integração destes modelos preditivos no âmbito da Indústria 4.0 permite operações industriais mais eficientes, fiáveis e sustentáveis.Recomenda-se trabalho futuro para melhorar a generalização do modelo, integrar dados em tempo real e expandir a aplicabilidade destas soluções em vários sectores industriais.Este trabalho tem como objetivo reduzir significativamente os custos de manutenção, minimizar o tempo de inatividade do equipamento e melhorar a eficiência operacional, apoiando assim o funcionamento sustentável e rentável das indústrias na Quarta Revolução Industrial.

Remote fault detection is essential for maintaining the performance and reliability of heavy-duty machines. This paper proposes a methodology for remote fault detection using a combination of the IEC 61499 standard and a Python-based runtime environment, DINASORE. The proposed methodology consists of two phases: (1) reference curve acquisition and (2) operation curve acquisition. In the first phase, reference curves are captured for each actuator in the machine when it is new. In the second phase, operation curves are captured for each actuator during machine operation. The operation curves are then compared to the reference curves to detect anomalies, using a combination of Dynamic Time Warp, Pearson Correlation, and Time Series Forest. The proposed methodology was evaluated using a real-world heavy-duty machine, and the results showed that it can be used to assist technicians in remote fault finding. The methodology is also scalable due to the modular nature of IEC 61499 and can be easily adapted to different types of machines or analysis techniques.

Abstract Designing and developing distributed cyber-physical production systems (CPPS) is a time-consuming, complex, and error-prone process. These systems are typically heterogeneous, i.e., they consist of multiple components implemented with different languages and development tools. One of the main problems nowadays in CPPS implementation is enabling security mechanisms by design while reducing the complexity and increasing the system’s maintainability. Adopting the IEC 61499 standard is an excellent approach to tackle these challenges by enabling the design, deployment, and management of CPPS in a model-based engineering methodology. We propose a method for CPPS design based on the IEC 61499 standard. The method allows designers to embed a bio-inspired anomaly-based host intrusion detection system (A-HIDS) in Edge devices. This A-HIDS is based on the incremental Dendritic Cell Algorithm (iDCA) and can analyze OPC UA network data exchanged between the Edge devices and detect attacks that target the CPPS’ Edge layer. This study’s findings have practical implications on the industrial security community by making novel contributions to the intrusion detection problem in CPPS considering immune-inspired solutions, and cost-effective security by design system implementation. According to the experimental data, the proposed solution can dramatically reduce design and code complexity while improving application maintainability and successfully detecting network attacks without negatively impacting the performance of the CPPS Edge devices.

Several factors have led to an increase in the focus on sustainable development. In this context, the concept of Circular Economy (CE) has gained tremendous momentum in research and implementation. Fitted into CE's concept, the Life Extension Strategies (LES) have been popular among industrial practitioners, regulators, policymakers, and academics from different industries due to the several benefits of these strategies. The general scope of this study is to approach LES within the industrial environment. The main goal is to understand how Remanufacturing and Refurbishment (R/R) strategies have been applied in different industry types. To achieve this goal, we carried out a systematic literature review whereby we captured some examples of R/R applications that demonstrate the potential of application to various industries (e.g., aerospace/aeronautics, energy power, and automotive industries) and equipment types (e.g., nuclear reactor, hydropower plant, and turbine blades). We also described some strategies to implement LES (e.g., economic analysis and life cycle assessment) and the different impacts achieved or expected (life extension, cost-saving, and efficiency enhancement). Moreover, we discussed three specific points concerning R/R application, the LES categorization, and the inexistence of a multi-dimensional framework to LES.

Industries generate and collect a huge amount of data about their processes. Usually such data contain relevant information, which can be used to monitor and analyze the processes, but also improve them, applying optimization techniques that allow to enhance different aspects, such as machine maintenance scheduling, product quality, use of resources and so on and so forth. The Digital Twin (DT) concept has been recently applied in the context of Industry 4.0, to exploit this data, leveraging advanced physical modelling, data analysis, Artificial Intelligence (AI) algorithms for optimization and prediction, which are two key concepts in the Industry 4.0 paradigm. Currently, the major challenge in this field is to make these techniques available for the industries and ensure their ease of use for the final users. This paper presents an holistic solution that combining two open-source software - i.e., REclaim oPtimization and simuLatIon Cooperation in digitAl twin (REPLICA) and Optimization Platform for Refurbishment and Re-manufacturing (OPR2) - provides a flexible, open and easy-to-use AI environment that allows data scientists to create, test, connect and deploy their algorithms and to optimize them. This work presents a first prototype of this solution, then describes how it has been tested and validated with real industry data and finally provides the results obtained with these tests.

This project aims to design a detection and pose estimation pipeline for objects, to be used in an industrial environment in a robotic grasping setting.

Over the past decades, software engineering has reached a level of maturity which entails great challenges in its education. Universities must prepare students to real-life challenges by offering courses to aid students in developing several vital skills which go beyond hard skills (e.g., communication skills and self-management). At the Faculty of Engineering of the University of Porto, a pioneering course, dubbed Project Management Laboratory, offers the proper environment for students to develop such skills by inviting industry to be closely involved in the education of the students. This course integrates practice and theory in a setting close to what the students will face when they move into industry. This paper reports on the experience, results, and benefits of this innovative course.

Cyber-Physical Production Systems (CPPS) are key enablers for industrial and economic growth. The introduction of the Internet of Things (IoT) in industrial processes represents a new revolution towards the Smart Manufacturing oncept and is usually designated as the 4thIndustrial Revolution. Despite the huge interest from the industry to innovate their production systems, in order to increase revenues at lower costs, the IoT concept is still immature and fuzzy, which increases security related risks in industrial systems. Facing this paradigm and, since CPPS have reached a level of complexity, where the human intervention for operation and control is becoming increasingly difficult, Smart Factories require autonomic methodologies for security management and self-healing. This paper presents an Intrusion Detection System (IDS) approach for CPPS, based on the deterministic Dendritic Cell Algorithm (dDCA). To evaluate the dDCA effectiveness, a testing dataset was generated, by implementing and injecting various attacks on a OPC UA based CPPS testbed. The results show that these attacks can be successfully detected using the dDCA.

Over the last years, robotics has increased its interest in learning human-like behaviors and activities. One of the most common actions searched, as well as one of the most fun to replicate, is the ability to play sports. This has been made possible with the steady increase of automated learning, encouraged by the tremendous developments in computational power and improved reinforcement learning (RL) algorithms.This paper implements a beginner Robot player for precision ball sports like boccia and bocce. A new simulated environment (PrecisionBall) is created, and a seven degree-of-freedom (DoF) robotic arm, is able to learn from scratch how to win the game and throw different types of balls towards the goal (the jack), using deep reinforcement learning. The environment is compliant with OpenAI Gym, using the MuJoCo realistic physics engine for a realistic simulation. A brief comparison of the convergence of different RL algorithms is performed. Several ball weights and various types of materials correspondent to bocce and boccia are tested, as well as different friction coefficients. Results show that the robot achieves a maximum success rate of 92.7% and mean of 75.7% for the best case. While learning to play these sports with the DDPG+HER algorithm, the robotic agent acquired some relevant skills that allowed it to win.

In today's industry, production processes are more oriented towards customer customization, demanding manufacturing plants to be increasingly flexible, where Human-Robot Collaboration (HRC) plays an important role. To fully take advantage of this collaboration, both robot and human need to perceive each others actions and intentions, operating accordingly. Thus, the typical collaborative environment that is nowadays monitored only for safety purposes needs to evolve into a more transparent, informative and attainable concept in order to give human-like perception to the robot.This paper proposes a voxel-based space monitoring approach in collaborative robotics environments, where distinct technologies are combined to form a labeled occupancy voxel-grid (LOG), i.e, a three-dimensional grid with labels for all the critical elements of the collaborative environment. A stereo vision camera is used to capture the supervised space in a point cloud, to then create an unlabeled voxel-grid. Making use of the RGB frames, both human and robot joint positions are located (using OpenPose and robot controller), pinpointing the positions of other significant elements in collaborative tasks as well. These positions are used to label the base voxel-grid. With the composition of the collaborative space provided in the grid, not only typical obstacle avoidance can be achieved, but also more advanced topics like predictive control or task recognition. Overall, this approach provides a much higher perception of the collaborative environment, enabling a more symbiotic relation between human and robot in collaborative robotics.

Absolute automation in certain industries, such as the automotive industry, has proven to be disadvantageous. Robots are fairly capable when performing tasks that are repetitive and demand precision. However, a hybrid solution comprised of the adaptability and resourcefulness of humans cooperating, in the same task, with the precision and efficiency of machines is the next step for automation. Manipulators, however, lack self-adaptability and true collaborative behaviour. And so, through the integration of vision systems, manipulators can perceive their environment and also understand complex interactions. In this paper, a vision-based collaborative proof-of-concept framework is proposed using the Kinect v2, a UR5 robotic manipulator and MATLAB. This framework implements 3 behavioural modes, 1) a Self-Adaptive mode for obstacle detection and avoidance, 2) a Collaborative mode for physical human-robot interaction and 3) a standby Safe mode. These modes are activated with recourse to gestures, by virtue of the body tracking and gesture recognition algorithm of the Kinect v2. Additionally, to allow self-recognition of the robot, the Region Growing segmentation is combined with the UR5's Forward Kinematics for precise, near real-time segmentation. Furthermore, self-adaptive reactive behaviour is implemented by using artificial repulsive action for the manipulator's end-effector. Reaction times were tested for all three modes, being that Collaborative and Safe mode would take up to 5 seconds to accomplish the movement, while Self-Adaptive mode could take up to 10 seconds between reactions.

Manufacturing companies are seeking forms of maximizing profits, where reduction of maintenance costs plays a critical part. Avoiding unexpected breakdowns while maintaining productivity is possible through continuously monitoring machine performance, predicting when and where a failure will occur. This allows not only to reduce downtime but also to apply the best maintenance strategy and assure production targets. In this paper, a Continuous Maintenance System to achieve this is proposed. This system joins a Predictive Maintenance module with optimization and simulation modules. The Predictive Maintenance module makes use of a Gradient Boosting Classifier to predict which machine component will fail and schedule its maintenance. The optimization module uses a Genetic Algorithm to find the throughput values that reveal the best balance between production and degradation rates, and therefore, changing maintenance schedules according to production targets and machine degradation. Finally, a statistical simulation model based on real data distribution was used to examine effects of a certain throughput and maintenance schedule for each machine. Several classifiers were tested for the predictor, comparing their performance. Also, 3 different scenarios of a parallel production line were used to evaluate the proposed system.

Transfer Learning aims at transferring knowledge from an already learned task to a different, but related task, in order to accelerate the learning process of the latter. This concept can be applied to manufacturing systems where process models that map process parameters into process quality are used to optimize the calibration phase of new unseen products at the shop-floor. However, these process models often require a great amount of experiments, which normally is costly and impractical for most manufacturing systems. The present work explores a Laser Seam Welding scenario with 3 different product variants where the problem is training one of the process models with a reduced amount of labeled data. Artificial Neural Networks (ANNs) were used to model these processes and Inductive Transfer Learning is then used to tackle the proposed problem. Ultimately, this approach was compared to traditional machine learning where no transfer occurs and a model is trained only using the small amount of labeled data. The results revealed that for all the Laser Seam Welding processes the trained models performed better when using Inductive Transfer.

The present paper details a novel methodology called Meta-Process Model that is able to generate new data-based models for manufacturing processes when no experimental data is available. For that purpose, the concept of Hyper-Models was used to create a higher level of abstraction of these manufacturing processes, along with a Statistical Shape Model (SSM) that is able to capture the modes of shape variations and build up a deformable model to generate new shapes. The main premise of the present work is to interpret a process model as a n-dimensional shape and use SSM to capture the variations among a set of different process models. This methodology is evaluated by using two already existing process models for a model generalization, from which a new process model is derived just with new, given process conditions. This new process model is then compared with a process model, which was independently estimated using real experimental data acquired under the same process conditions. The results show that a previously nonexistent process model that captures the dynamics of the real process can be generated, even when there's no experimental data and only the new process conditions are available.

Cyber-Physical Production Systems (CPPS) are the key enabling for industrial businesses and economic growth. With the introduction of the Internet of Things (IoT) in the manufacturing environment, CPPS have a huge potential in terms of new business opportunities, mostly known as the 4thIndustrial Revolution. The paper presents a CPPS architecture achieved within the European R&D project SelSus, which is validated in an industrial human-machine collaborative use case scenario. This scenario is composed by a set of sequential gripping operations between a person and a robotic arm, where both sensing and actuation devices are virtualized using the SmartComponent concept. The developed prototype demonstrated that the SelSus CPPS enabled self-adaptation in industrial equipment, in order to help the human operator under high levels of stress and fatigue.

This work was developed in the context of European funded Project SelSus. Industrial clouds are heavily sensor based and Cloud Manufacturing Service frameworks are mostly grounded in the adoption of Internet of Things and Wireless Sensor Networks technologies. The SelSus framework combines both an Industrial Sensor Cloud and a Cyber Physical Production System. The Industrial Sensor Cloud supports the Cyber Physical Production System, by providing computational power, as well as internal coordination and control, and external access to realize intelligent monitoring and control. The SelSus framework combines embedded systems, networks, sensors and actuators and control algorithms in a seamless manner. Our goal is to have a Flexible Sensor Integration solution that allows rapid graphical development of interpreters of raw data packets in the Cloud and its deployment for embedded execution at the Wireless Sensor Network gateway level for automatic data acquisition. This paper describes such a technology and demonstrates its feasibility, where a match between a graphically developed interpreter and the received messages from the Wireless Sensor Network in US-ASCII is made, and the integration of such sensors is made into the system.

Sensor data is extremely important to monitor machines at the shop-floor level and its environmental surrounding conditions for condition-based monitoring, machine diagnosis and process adaptation to new requirements. Based on the described scope, self-diagnostics and self-organizing capabilities are core functionalities of any Industrial Wireless Sensor Network (IWSN). In the present work, a simulated case study was developed with the main intent of validating techniques implemented for sensor data diagnosis of error detection and equipment failure. The scenarios explored try to mimic some common situations of a manufacturing environment when dealing with WSNs, where a piece of sensor equipment suddenly stops working or an unpredictable change in the environment leads to faulty data readings. This paper introduces Castalia and describes how it was used to simulate a direct application of an Optical Metrology System on an industrial Resistance Spot Welding process, which is composed of a camera and several luminosity sensors. More specifically, a sensor data validation module was proposed, implemented and used to extend Castalia functionalities.

Sensor networks that consist of a variety of sensor nodes, with capability for distributed storage and analysis, interoperable and delay-tolerant communication, will pave the way for a truly scalable network of sensors which will support adaptable plug-and-produce assembly stations. The concept of Sensor Cloud has emerged as the cornerstone for enabling the integration of nearly real time data sources into Service Oriented Architectures and as one of the enablers for Reconfigurable Manufacturing Systems. Hitherto there are still some challenges that need to be tackled, namely the on-the-fly instantiation and update of services at the system level and the dynamic (re)organization of the services created. This paper presents the first steps in the development of a framework (taking advantage of several technologies like UPnP, OSGi and iPOJO) to address these challenges and make Sensor Clouds a reality in the shop floor. The results from the first implementation reveal that the performance of the system is linear in terms of scalability, and from these scalability tests, we proved the framework's robustness and consistent responsiveness.

Sensors always played a significant role on the industrial domain, since monitoring the current machine's process state is notoriously an advantage for shop-floor analysis, and consequently, to rapidly take action according to the production system demands [3, 6, 7, 8]. The I-RAMP3European Project explores exactly these demands, and proposes new approaches to efficiently address some of the nowadays difficulties of the European Industry. The Smart Sensor technology is explored in the I-RAMP3Project using the NETwork-enabled DEVice (NETDEV) concept, as a logical entity for equipment encapsulation with high level of communication capabilities and intelligent functionalities. Therefore, not only how the NETDEV concept is implemented, but also how to use sensors is explored in the present paper, by means of UPnP Technology, for communication extensibility, and PlugThings Framework for easy sensor integration and complexity addition. Moreover, the importance of sensors based on the context of the I-RAMP3is explored, discussing some trends and possible steps to be taken in conventional production towards the next generation of Smart Manufacturing Systems.

The project XPRESS introduces a completely new scalable concept of an expertonic networked factory, which is composed by a coordinated team of specialized autonomous entities (intelligent production units), each knowing how to do a certain process optimally. This paper looks in particular to the transport intelligent unit implementation, which is responsible for the transportation of components at the shop-floor. An architecture based in a multi-agent approach that uses automatic guided vehicle expertons to perform this task is suggested. Additionally, it is presented an implementation of the automatic guided vehicle transport intelligent unit that uses the .NET framework integrated with the expertonic framework.

The XPRESS project defined a new concept of intelligent factory to improve the flexibility of the manucfacturing lines. One of the problems the management layer of this system faces is the decision regarding the composition of new production lines to realize a product. To solve these issues, this work focuses on developing methods to assist the engineer decision by eliminating undesired equipments combinations from the elegible configurations list. The result was a library of optimization functions for combinatorial problems and it is composed by three methods, each corresponding to a different stage of the decision process. This work also analyzes the impact of the problem features in the produced implementation.

This video shows a brief overview over Neptus, a command and control infrastructure for heterogeneous teams of autonomous vehicles. Having different types of vehicles at our laboratory and from our partners, there was an increasing need to create a common infrastructure to all these systems. Additionally, a tool to support the entire mission life cycle (planning, execution, review and dissemination) was lacking. Neptus was created to provide vehicle independence and seamless inter-systems communications. Currently, Neptus has been already tested with remotely operated vehicles, autonomous underwater vehicles, unmanned air vehicles, autonomous surface vehicles and wireless sensor networks. Some of these systems were operated simultaneously by various operating consoles that were sharing the same communication infrastructure. The received data was being relayed to a Web server that allowed for the real-time mission following by using a common Web browser.

This paper presents the Mission Review and Analysis module of a C3I (Command, Control, Communication and Information) infrastructure - the Neptus framework. This is a mix-initiative environment that it's being developed in the Underwater Systems and Technology Laboratory (USTL/LSTS) with the goal to support the coordinated operation of heterogeneous teams of vehicles. This includes autonomous and remotely operated underwater, surface, land, and air vehicles and people. People perform a fundamental role, not only in the case of remotely operated vehicles, but also with autonomous vehicles where mix-initiative operation is a requirement. The Neptus is very modular and will be heavily based on services with a distributed architecture. This paper focus mainly on the mission review and analysis where the data collected in a mission is prepared for analysis.

The C3I (command, control, communication and information) Neptus framework which is being developed at the Underwater Systems and Technology Laboratory (USTL/LSTS) is presented. Neptus is a modular mixed initiative framework (human operators in the control loop) for the operation of heterogeneous teams of vehicles such as autonomous and remotely operated underwater, surface, land, and air vehicles. Neptus is composed of mission and vehicle planning, supervision, and post-mission analysis modules which are provided as services across a network. This paper focus mainly on the mission definition module with the presentation of MDL - a XML based language for mission definition

This paper describes a systems engineering approach to the design of an integrated decision support system, IDES, for the planning and control of ATMA, a Portuguese textile company. The paper presents the overall process emphasizing the methodological framework and the main results. Systems engineering methods were crucial in the development of a methodological framework adapted to the specificities of the Portuguese industry. This framework proved invaluable in the process of mapping a conceptual planning and control architecture based on dynamical optimization onto the design of the IDES according to the recommendations of the strategic plan of ATMA. Concepts and techniques from control systems, dynamic optimization, software engineering, enterprise modeling and anthropocentric systems were used in this process.