Title Promoter Affiliations Abstract "Assuring the safety of autonomous systems through human-centered executable assurance cases" "Davy Pissoort" "Distributed and Secure Software (DistriNet), Waves: Core Research and Engineering (WaveCore)" "While autonomous systems offer tremendous possibilities, they also come with major safety challenges. After all, human safety is central to every autonomous system. Existing safety assurance approaches and standards have been developed primarily for systems where a human can take over in the case of emergency and do not extend to autonomous systems. The combination of executable safety assurance cases and digital twins has tremendous potential as a solid safety framework for autonomous systems. However, the transition from classic static-safety assurance during design-time to dynamic safety guaranteed by the system itself without further human intervention is a big step to take. Therefore, this PhD proposal will focus on the concept of human-centered executable safety cases as an important intermediate step. More specifically, we will investigate the following research hypothesis: can we enable the safety engineer to dynamically re-evaluate the safety claims, arguments and assumptions underlying the safety assurance case and select the appropriate course of action by combining the concepts of executable safety assurance cases and digital twins. This research will create fundamental knowledge on dynamic risk management, necessary to assure the safety of modern high-tech, software-driven autonomous systems. Additionally, the results of this research supports the transition towards  increased (optimal) use of smart devices  and safe intelligent transport systems. " "SaFeR: European Training Network on Safety of Autonomous Systems" "Jeroen Boydens" "Distributed and Secure Software (DistriNet), Waves: Core Research and Engineering (WaveCore)" "To realize and fully exploit the positive potential of autonomy, community needs to have ways of making such systems safe. The specific safety challenges of autonomous systems and the technologies that enable autonomy are not adequately addressed by current safety management practices and standards. It is clear that autonomous systems can introduce many new paths to accidents, and that autonomous system technologies may not be practical to analyse adequately using accepted current practice. As an example, the current major safety standard for cars – ISO 26262 – in fact relies on a driver being there and being under control. New methodologies need to be developed and young engineers need to be trained on how to (1) analyze the possible risks and hazards related to autonomous systems, (2) how to make autonomous systems safety-by-design, and (3) how to come to an evidence-based safety-case for autonomous systems. " "Adaptive Human Operator Interaction with Autonomous Systems (AHOI)." "Dirk Van Rooy" "Antwerp Maritime Academy, IMEC, Product development" "In autonomous systems, human interaction is key. While machines handle uncertain scenarios, they often lack in ethical considerations. AI systems, being ""black boxes,"" make decisions that are not transparent. This lack of transparency can challenge human operators, especially in risky situations, leading to monitoring difficulties and ethical issues. Human biases can also impact this interaction, resulting in overconfidence or dismissal of AI recommendations, thus breaking down trust in the system. Addressing this, the Adaptive Human Operator Interaction with Autonomous Systems (AHOI) consortium aims to understand how explainability and trust in AI shape each other. This team of diverse researchers focuses on trust in the context of maritime autonomous navigation. The research includes: 1. Developing a robust autonomous navigation system for dynamic maritime environments, using advanced machine learning for unknown settings. 2. Studying how explainable AI (XAI) affects operators' trust and decision-making, enhancing understanding of AI decisions for operators with varying expertise. 3. Investigating the interplay between human biases and transparency in human-machine interaction, identifying optimal collaboration points. 4. Designing an advanced human-machine interface (HMI) that offers insights into AI decision processes, tailored to user experience levels. 5. Using XAI and visualization software to create a dynamic, interactive HMI that adapts explanations based on user feedback, fostering continuous learning. This comprehensive approach aims for a system that is both robust and transparent, facilitating efficient human-machine collaboration in maritime navigation. Though focused on maritime navigation, the findings have broader applications in defense, like mine hunting, surveillance, UAV, and UGV operations. In AHOI, iMec will research AI and XAI, UA and AMA will study human operator biases, and MAHI will focus on autonomous vessels' situational awareness and HMI design." "SAFETEE - Towards a computing platform for safe autonomous systems" "Tom Holvoet" "Waves: Core Research and Engineering (WaveCore), Distributed and Secure Software (DistriNet)" "Autonomous systems (cars, drones, off-road vehicles, etc.) can bring tremendous advancement for people and businesses. A crucial concern for such systems is safety. While most studies today focus on reliable sensors and perception, and actuators and control, this project aims for a vertical, in-depth study of a full-stack computing platform. Starting from a systematic approach for requirements gathering for the computing platform, the project will study (1) approaches for enhanced reliable hardware under harsh environment conditions, (2) error recovery algorithms e.g. in case of bit flips, (3) a safe and secure reactive software execution platform with formally verified scheduling mechanisms and latency guarantees, and (4) on top an innovative decision making approach for autonomous systems (the ‘brain’) that verifiably guarantees safe decision making. This unique combination of interacting aspects of a safety-aware platform is highly innovative and appealing to both research and industry." "Safer Autonomous Systems" "Davy Pissoort" "Waves: Core Research and Engineering (WaveCore), Distributed and Secure Software (DistriNet), Research Unit KU Leuven Centre for IT & IP Law (CiTiP)" "Autonomous systems offer humankind tremendous opportunities, like freeing us from mundane tasks, carrying out risky procedures and generally giving us more time to enjoy the things we like doing. However, we lack trust in many forms of autonomous systems: partly this is human nature, but primarily because these systems, such as self-driving cars, have not demonstrated their safety credentials. Only by making these systems safer can we expect their widespread acceptance. The Safer Autonomous Systems (SAS) ETN is about getting people to trust these systems by making the systems safer. In order to achieve this objective and to train a group of highly skilled, responsible, future innovators, we will bring together 15 early-stage researchers (ESRs) to investigate new forms of system-safety engineering, dependability engineering, fault-tolerant and failsafe hardware/software design, model-based safety analysis, safety-assurance case development, cyber-security, as well as legal and ethical aspects. SAS will actively research the development of safer autonomous systems at multi-nationals like Bosch, but it also wants to stimulate the development of new safety designs, modelling and assurance techniques by involving the ESRs in SMEs and, potentially, their own start-ups. To help the ESRs put what they have learned during their research and S/T training into practise in their future careers, they will also receive soft-skills training to help them communicate effectively at all levels and become sought-after recruits. SAS is closely aligned with the high-priority areas of the EU, addressing many Horizon 2020 thematics, e.g., Industrial Leadership (Advanced manufacturing and processing), Societal Challenges (Smart, green and integrated transport; Secure, clean and efficient energy) and Excellent Science. But the most important output of SAS will be 15 well qualified people who have been trained to tackle many of the problems now being faced by European industry." "Spline-Based Motion Planning for Autonomous Mechatronic Systems" "Goele Pipeleers" "Production Engineering, Machine Design and Automation (PMA) Section, Robotics, Automation and Mechatronics (RAM)" "In modern industry there is an ever lasting quest to obtain a higher productivity at lower costs. Introducing autonomous motion systems provides a huge potential to make progress in this quest, since they allow drastically improving the efficiency of various industrial processes and tasks, including harvesting fields with tractors, order picking in eCommerce warehouses and container transport through ports. In addition, automation may also be beneficial in our everyday life, for example in applications such as robotic surgery, personal assistance robots or parcel delivery by drones.When studying the case of autonomous systems moving around in industrial environments, it becomes clear that the currently existing set-ups are conservative and require strongly conditioned environments, limiting their capabilities. As a consequence, there is a growing interest in more flexible techniques, allowing free motion of autonomous systems through unconditioned environments.This thesis proposes a novel spline-based motion planning approach to compute trajectories that steer holonomic and nonholonomic systems trough uncertain and dynamic environments while taking into account the kinematic vehicle limits and avoiding collision with all obstacles. In order to deal with uncertainties and changes in the environment, the motion planning problem is solved online, during the vehicle motion. By exploiting spline properties, a small-scale optimization problem is obtained, allowing a low solving time and therefore a swift reaction to changes. In addition, these spline properties allow obtaining guaranteed constraint satisfaction at all time instants.In vast environments containing a multitude of obstacles, solving the complete motion planning problem at once is very complicated. In addition, obstacles moving near the goal position influence the complete computed trajectories, while due to uncertainties it is not useful to already take these obstacles into account at the start of the system's motion. Therefore, in the case of a vast environment, a scheduler combines the presented spline-based method with a global path planner in order to split the complete motion planning problem over a series of subproblems that are easier to solve. The resulting approach outperforms most existing ones since it allows substantial deviation from the global path and fluently deals with uncertainties and changes in the environment.When considering the case of motion planning for CNC machines, trajectories must be computed that move a machine tool as fast as possible along the contour of the desired workpiece. For this case, a similar approach can be applied as in a vast environment: instead of solving the complete problem at once, only the trajectories for a specified number of segments, building up the contour, are computed simultaneously. The problem formulation includes the kinematic limits of the tool, together with the process limits and the given tolerance. This allows computing feasible trajectories that exploit the tolerance to cut corners, reducing the machining time. In addition, to avoid decoupling between subsequent sets of segments and further reduce the required machining time, the problem is solved with a moving horizon.The potential of the designed approach is demonstrated with extensive numerical simulations. In addition, experiments are performed on respectively the KUKA youBot and a 3-axis micro-milling machine. OMG-tools, a user-friendly open-source motion planning toolbox, is extended to include all methods presented in this thesis. The toolbox makes it easy to set-up, solve and simulate problems by using the proposed spline-based motion planning approach." "Safe and Dependable Autonomous Systems - Formal specification and verification" "Tom Holvoet" "Waves: Core Research and Engineering (WaveCore), Distributed and Secure Software (DistriNet)" "This PhD project is part of a research track that studies and develops a computing platform for trustworthy autonomous systems with provable characteristics for safety and resilience. In particular, this PhD project will study and assess the application STPA to the HARA of autonomous systems on three facets: (1) the suitability of STPA for determining and modelling the substantial sets of safety requirements for autonomous systems; (2) concepts and techniques for verifying the completeness and consistency of the safety requirements; (3) its ability to account for self-adaptive autonomous systems (e.g. due to machine learning and AI) - since we need to scope this proposal, we focus such adaptations to basic parameter updates. A use case of an AMR (Autonomous Mobile Robot) will be used to validate the research." "Software for cooperative autonomous systems" "Tom Holvoet" "Distributed and Secure Software (DistriNet)" "Developing software for cooperative autonomous systems is challenging, especially for safety-critical systems such as UAVs. A problem with current approaches in the design of UAV software is handling conflicting goals. Currently, there is no way to detect/avoid conflicting goals in the UAV software in a consistent and predictable manner. Component-based design and development can bring inherent reliability to an airborne system, however, these are rarely considered in current UAV component development. No work is available surrounding reliability by design of the integration of flight critical systems.In this research we want to develop a methodology and a software architecture which is safety- and autonomy-oriented and can provide guarantees regarding the expected behaviour of applications created according to the methodology. The first goal is to define a component model that works within the constraints of real-time software. The second goal is to create a model for autonomous software entities, taking safety guarantees into account. The third goal of this task is to define a safety specification language, which is expressive enough to represent the high level safety requirements, but also formal enough to allow the verification of a safety specification written in such language." "Conditioned harsh outdoor environment for perception systems of autonomous applications (CAVE)." "Walter Daems" "Co-Design of Cyber-Physical Systems (Cosys-Lab)" "Autonomous mobile systems might fail for many reasons, but one of them is when the harshness of the environment increases. It is difficult for OEMs, integrators, sensor and hardware components providers to design a robust autonomous mobile system based on traditional testing methods. Especially perception systems are challenged in realistic and relevant harsh conditions (e.g. rain, fog, direct sunlight). Currently, testing of perception systems is done by waiting for these conditions to happen in real-life – which can easily cost weeks of waiting. When an update is done on the hardware of the perception system (e.g. a coating on the lens is added) the exact same test is needed to verify an improvement. However, in real-life this exact same harsh condition cannot be reproduced. So, there is a need for a modular, validated testing facility that allows controllable and measurable conditions, to enable repeatable and controlled harsh conditions. CAVE_INFRA aims to develop a fixed perception test facility which can control and measure rain, snow, fog, illumination, dust and debris conditions, including its digital twin and a real-life validation. We aim to provide the following services: i) Sensing hardware (incl coatings/cleaning systems) and software performance evaluation in harsh conditions, including benchmarking to support sensor selection ii) Harsh condition model and/or sensor model derivation iii) Training or validation of AI models for objects / human detection and pose estimation iv) Degradation tests in harsh conditions v) Generate test data and scenarios that can be used for driving out own research but also for certification purposes and discussions with certification bodies such as TuV. To produce the harsh conditions in realistic scenarios, there are different actuation systems foreseen to respectively actuate the perception system under test, the target objects to be detected, and some of the generated conditions such as diverse illumination systems to create dynamic contrast." "Adaptation of task execution strategies in dynamic and uncertain environments by cooperation between networked autonomous systems." "Pieter Simoens" "Department of Information technology" "Application developers cannot anticipate at design time the scale and complexity of real-life environments where networked autonomous systems will be deployed. The goal of this project is to enable such systems to adapt task execution to the runtime context through resource-aware negotiation and invocation of capabilities provided by other systems in the environment."