Diffusion models have shown great potential in generating trajectory plans for agents in environments with unknown dynamics. However, such models provide no safety guarantees. In this work, we focus on risk-aware planning with respect to safety constraints and introduce a probabilistically risk-aware variant of Diffuser (PRA-Diffuser). The diffusion model initially learns a distribution over trajectories that may or may not be unsafe. We then fine-tune this model to reduce the probability of sampling such unsafe trajectories. We analyze the proposed solution and introduce a provable lower bound on risk of safety violation leveraging concentration inequalities for conditional Value-at-Risk. Our approach can be applied to models that have been pre-trained, potentially from datasets containing unsafe trajectories. Our empirical results demonstrate that our approach significantly reduces unsafe trajectories generated by the diffusion model across multiple environments.

We study multi-agent reinforcement learning (MARL) under temporally extended Signal Temporal Logic(STL) objectives, which require reasoning over both long-horizon dynamics and inter-agent relations. Wepropose TD-MAT, a transformer-based architecture with multivariate positional encodings, causal temporalmasking, and a decomposed reward based on arithmetic–geometric mean robustness with variance regularization. Experiments on coordination tasks ranging from unstructured multi-objective problems to strict temporalsequencing show that TD-MAT learns effective long-term behaviors and generalizes to heterogeneous agentsettings. Ablation studies highlight the necessity of temporal masking, positional encodings, and reward decomposition, while comparisons to MAPPO, RMAPPO, and MAT reveal that transformers provide the greatestbenefit on unstructured, long-horizon tasks.

Continuous integration and delivery (CI/CD) have transformed software development by reducing delivery time, improving product quality, and giving enterprises a competitive advantage. However, large-scale projects confront difficulties in giving fast feedback to developers due to large test suites, resulting in longer testing cycles and lower productivity. Traditional regression testing methods struggle to find a balance between efficacy and efficiency, demanding advanced approaches. This study investigates the use of machine learning (ML), specifically Neural Network and Random Forest models, to choose test cases based on source code changes, commit messages, and change file path in order to offer developers with faster feedback. The study investigates the predicted accuracy of ML models using a large industrial dataset from a telecom company, which included 15 million test executions over 15 months. The results show that Random Forest outperforms Neural Network models in test case selection, with up to 97% accuracy achieved. Real-time evaluations conducted over a month show significant savings in test executions (88 % -90 %) and testing time (44 % -74%) across multiple regression testing activities, illustrating the potential of ML-driven techniques to optimize CI/CD pipelines and increase developer productivity.

With the emergence of Industry 4.0, comes an increasing need for multi-robot coordination and communication to efficiently complete joint tasks. A critical technology is the fifth generation (5G) mobile network, which enables cloud-controlled robots to execute tasks with differentiated quality-of-service (QoS) features. While there has been significant research on multi-robot planning, the integration with the capabilities of realistic network systems has been limited. In this paper, we introduce RoboPlan5G, a framework for 5G-aware robot planning. We propose a joint state-search model that includes task planning coordination in conjunction with 5G physical resource block (PRB) allocation. This process ensures efficient usage of the limited indoor 5G spectrum and effective service performance, while still aiming for the completion of tasks in the shortest possible time frame. Scenarios are generated in an Industry 4.0 environment, where the planner is shown to decrease the required 5G spectrum allocation by 50%, and on average improve the plan quality by 45% while maintaining a small computation time.

With the emergence of Industry 4.0, comes an increasing need for multi-robot coordination and communication to efficiently complete joint tasks. A critical technology is the fifth generation (5G) mobile network, which enables multiple robots to execute control tasks with differentiated quality-of-service (QoS) features. However, there has been limited analysis of the impact of real 5G capabilities on multi-agent robot planning problems. In this paper, we provide a review of robot planning algorithms suitable for industrial use-cases, which consider communication aspects in the planning formulation. The paper is further positioned to identify gaps in existing state of the art within communication-aware planning. This is followed by an analysis of key challenges to be targeted at the intersection of 5G, Industry 4.0 and multi-agent robot planning. This analysis is strategically important and would prove useful to academic researchers and industry experts focusing on deployment of robots in industrial settings.

This is Part II of a double-part special issue on distributed learning over wireless edge networks. This two-part special issue features papers dealing with two main research challenges: optimization of wireless network performance for efficient implementation of distributed learning in wireless networks, and distributed learning for solving communication problems and optimizing network performance. The accepted papers in this special issue have been grouped into three topics: 1) network optimization for federated learning (FL), 2) network optimization for other distributed learning methods, and 3) distributed reinforcement learning (RL) for wireless network optimization. In Part I (vol. 39, no. 12, Dec. 2021), the focus is on the first cluster (network optimization for FL). The focus of Part II is on the second and third clusters (network optimization for other distributed learning methods and RL for wireless network optimization). The readers are referred to Part I for an overview paper [A1] by the team of guest editors where a comprehensive study of how distributed learning can be efficiently deployed over wireless edge networks is provided. The contributions made by the papers in Part II are summarized as follows.

Analyzing massive amounts of data using complex machine learning models requires significant computational resources. The conventional approach to such problems involves centralizing training data and inference processes in the cloud, i.e., in data centers. However, with the proliferation of mobile devices and increasing application of the Internet-of-Things (IoT) paradigm, very large amounts of data are collected at the edges of wireless networks, and due to privacy constraints and limited communication resources, it is undesirable or impractical to upload this data from mobile devices to the cloud for centralized learning. This problem can be solved by distributed learning at the network edge, by which edge devices collaboratively train a shared learning model using real-time mobile data. The avoidance of raw-data uploading not only helps to preserve privacy but may also alleviate network-traffic congestion and minimize latency. With that said, distributed training still requires a substantial amount of information exchange between devices and edge servers over wireless links. In the process, wireless impairments such as noise, interference, and imperfect knowledge of channel states can significantly slow down distributed learning (e.g., convergence speed) and degrades its performance (e.g., learning accuracy). This makes it crucial to optimize wireless network performance so as to support the efficient deployment of distributed learning algorithms. On the other hand, distributed learning algorithms provide a powerful tool-set for solving complex problems in wireless communication and networking. One important framework, called federated learning (FL), enables users to collaboratively learn a shared model while helping to preserve local data privacy. The application of FL can endow edge devices with capabilities of user behavior prediction, user identification, and wireless environment analysis. As another example, distributed reinforcement learning is capable of leveraging distributed computation power and data to solve complex optimization and control problems that arise in various use cases, such as network control, user clustering, resource management, and interference alignment. To cover this paradigm of distributed learning over wireless networks, this two-part Special Issue features papers dealing with two main research challenges: a) optimization of wireless network performance for efficient implementation of distributed learning in wireless networks, and b) distributed learning for solving communication problems and optimizing network performance.

The next-generation of wireless networks will enable many machine learning (ML) tools and applications to efficiently analyze various types of data collected by edge devices for inference, autonomy, and decision making purposes. However, due to resource constraints, delay limitations, and privacy challenges, edge devices cannot offload their entire collected datasets to a cloud server for centrally training their ML models or inference purposes. To overcome these challenges, distributed learning and inference techniques have been proposed as a means to enable edge devices to collaboratively train ML models without raw data exchanges, thus reducing the communication overhead and latency as well as improving data privacy. However, deploying distributed learning over wireless networks faces several challenges including the uncertain wireless environment (e.g., dynamic channel and interference), limited wireless resources (e.g., transmit power and radio spectrum), and hardware resources (e.g., computational power). This paper provides a comprehensive study of how distributed learning can be efficiently and effectively deployed over wireless edge networks. We present a detailed overview of several emerging distributed learning paradigms, including federated learning, federated distillation, distributed inference, and multi-agent reinforcement learning. For each learning framework, we first introduce the motivation for deploying it over wireless networks. Then, we present a detailed literature review on the use of communication techniques for its efficient deployment. We then introduce an illustrative example to show how to optimize wireless networks to improve its performance. Finally, we introduce future research opportunities. In a nutshell, this paper provides a holistic set of guidelines on how to deploy a broad range of distributed learning frameworks over real-world wireless communication networks.

Realization of industry-scale, goal-driven, autonomous systems with AI planning technology faces several challenges: flexibly specifying planning goal states in varying situations, synthesizing plans in large state spaces, re-planning in dynamic situations, and facilitating humans to supervise, give feedback and intervene. In this paper, we present Intent-driven Strategic Tactical Planning (ISTP) to address these challenges. We demonstrate its efficacy through its application for radio base station inspection across several locations using drones. The inspection task involves capturing images, thermal images or signal measurements - called knowledge-objects - of various components of the base stations for downstream processing. In the ISTP approach, an operator indicates her goals by flying the drone to different components of interest. These goals are generalized to capture the intent of the operator, which are then instantiated in new situations to generate goals dynamically. Towards planning and re-planning in large state spaces to achieve these goals efficiently, we extend the Strategic-Tactical Planning paradigm. All the components of ISTP are integrated in an intuitive UI and demonstrated through a real life use-case built on the UNITY simulator platform.

With the continuous progress of radio technology it is becoming evident that the cellular network will support advanced cellular Vehicle-to-Everything (C-V2X) services, which require connectivity that can provide low latency, high data throughput, high reliability and broad coverage. Before these services are fully deployed in reality, and due to ethical and safety reasons, we turn to simulation to study the support of the cellular network for such services. Therefore, in this paper, we propose a realistic simulation framework for urban C-V2X usecases. The framework consists of an urban mobility simulator, which can accurately model urban traffic, and a network simulator capable for simulating at least the LTE radio equipment and the network protocol stacks in the urban area where the C-V2X service will be offered. We validate the simulation framework by performing a feasibility assessment of a vehicle teleoperation usecase in one urban area of the city of Stockholm.

Scheduling is needed in a multitude of smart city domains, such as public transportation, logistics, organization of events and parts delivery in assembly lines. In this paper, we present the architecture and implementation of PickUp Planner, a service that enables scheduling and dynamic routing of connected vehicles for picking up number of users distributed in the city. The service dynamically reconfigures the routes of the vehicles based on real-time traffic condition information, which in turn is gathered from the road infrastructure and other connected vehicles registered to the service. The crux of the service is the usage of a novel vehicle routing and scheduling algorithm that combines clustering algorithms and Answer Set Programming. Our evaluation shows that the algorithm decreases the total distance traveled by the vehicles, and therefore reduces fuel consumption and emissions from the vehicles. We also conclude that Answer Set Programming solvers may not be the best choice for scenarios in which there is a high number of users to be served frequently.

5G networks will be a key enabler for the Internet of Things by providing a platform to connect a massive number of devices with heterogeneous sets of network quality requirements. In this environment, 5G network operators will have to solve the complex challenge of managing network services for diverse customer sectors (such as automotive, health or energy) with different requirements throughout their lifecycle. In this paper, we present current state of our work on automating part of the network service lifecycle management using knowledge management- and decision support techniques. We also present our ongoing implementation steps for such management function in 5G networks.

Adoption of virtualization technology has been limited in industrial automation due to unavailability of mature solutions, and strict timing requirements of control systems. However, current advancement in Virtual Monitoring Machine, multicore technology, virtualization extension and network virtualization has led to increased interest of virtualization in industrial automation. So far, many related research are focused on maximizing CPU and I/O utilization, and optimization applicable to soft realtime systems (i.e., outside industrial automation domain), e.g., multimedia applications. In this research, we make use of QoS for CPU, memory and network bandwidth in pursuit of high speed and predictability on a distributed multicore platform which is constructed entirely from open source products. We evaluate the platform for latency and jitter, network throughput and CPU computation load. Finally, we analyze the result for applicability in industrial control domain.

With multicore controllers becoming available for industrial automation applications, new tools and algorithms to compute efficient partitioning and scheduling solutions for control applications need to be developed. Optimizing the deployment and the schedule of a set of Function Block Diagrams on a parallel architecture are both NP hard. Additionally, control engineers need help to shift from the single core towards the multicore paradigm. By taking advantage of the parallelism inside the control applications it is effectively possible to decrease the finish times of the applications which enables to decrease their cycle times and improve the quality of service of the controller processes. This paper presents a practical solution to this problem that consists in a framework, called PASA, designed for partitioning and scheduling control applications modeled as function block diagrams. It enables new algorithms tailored to solve these optimization problems. This paper presents an extension of list-based DAG scheduling algorithms designed to compute a deployment and schedule for several control applications with different cycle times. The different variants of this algorithm are compared against each other as well as against some other existing solutions on a set of randomly generated examples.

Virtualization is a promising solution to develop complex embedded systems with real-time requirements. This paper discusses the current state-of-the-art in virtualization technologies, with a particular focus on solutions for embedded real-time systems. Several such solutions have been developed over the past decade, and in this paper we give an overview of the more well known technologies and we provide a comparative assessment of key virtualization techniques available in these solutions. Gaps and lacking pieces are identified and further development and research is suggested.

In the last decade, a large number of different software component models have been developed, with different aims and using different principles and technologies. This has resulted in a number of models which have many similarities, but also principal differences, and in many cases unclear concepts. Component-based development has not succeeded in providing standard principles, as has, for example, object-oriented development. In order to increase the understanding of the concepts and to differentiate component models more easily, this paper identifies, discusses, and characterizes fundamental principles of component models and provides a Component Model Classification Framework based on these principles. Further, the paper classifies a large number of component models using this framework.

In this paper, we present a tool-chain for the REMES language, which can be used for the construction and analysis of embedded system behavioral models. The tool-chain consists of the following tools: (i) a REMES editor for modeling behaviors of embedded components, (ii) a REMES simulator to test timing and resource behavior prior to formal analysis, and (iii) an automated transformation from REMES to Priced Timed Automata, needed for formal analysis.

There is no clear distinction between service-oriented systems (SOS) and component-based systems (CBS). However, there are several characteristics that could let one consider SOS as a step further from CBS. In this paper, we discuss the general features of CBS and SOS, while accounting for behavioral modeling in the language called REMES. First, we present REMES in the context of CBS modeling, and then we show how it can become suitable for SOS. We also discuss the relation between our model and the current state of the art.

In this paper, we introduce the model REMES for formal modeling and analysis of embedded resources such as storage,energy, communication, and computation. The model is a state-machine based behavioral language with support for hierarchical modeling, resource annotations, continuous time, and notions of explicit entry and exit points that make it suitable for component-based modeling of embedded systems.The analysis of REMES-based systems is centered around a weighted sum in which the variables represent the amounts of consumed resources. We describe a number of important resource related analysis problems, including feasibility, trade-off, and optimal resource-utilization analysis.To formalize these problems and provide a basis for rigorous analysis, we show how to analyze REMES models using the framework of priced timed automata and weighted CTL. To illustrate the approach, we describe a case study in which it has been applied to model and analyze resource usage of a temperature control system.

The conflicting requirements of real-time embedded systems, e.g. minimizing memory usage while still ensuring that all deadlines are met at run-time, require rigorous analysis of the system's resource consumption, starting at early design stages. In this paper, we glance through several representative frameworks that model and estimate resource usage of embedded systems, pointing out advantages and limitations. In the end, we describe our own view on how to model and carry out formal analysis of embedded resources, along with developing the system.

In this paper we propose to use components for managing the increasing complexity in modern vehicular systems. Compared to other approaches, the distinguishing feature of our work is using and benefiting from components throughout the development process from early design to development and deployment, and an explicit separation of concerns at different levels of granularity. Based on the elaboration of the specifics of vehicular systems (resource constraints, real-time requirements, hard demands on reliability), the paper identifies concerns that need to be addressed by a component model for this domain, and describes a realization of such a component model.