About the project

Objective
This project advocates a paradigm shift in the development of CPS by proposing a secure-by-construction controller synthesis scheme that considers security properties simultaneously with safety ones during the design phase. To successfully overcome the design challenges encountered in large-scale CPS under complex security requirements, we aim to develop a compositional and automated secure-by-construction design process based on a cross-disciplinary approach combining theoretical techniques from computer science (e.g. assume-guarantee rules) with those from control theory (e.g. small-gain theorems). This project aims to bring a potential solution to the fundamental security issue for the smart society vision by enabling cost-efficient and reliable design for CPS with formal security guarantees.

Background
Cyber-physical systems (CPS) are the technological backbone of the increasingly interconnected and smart world where design faults or security vulnerabilities can be catastrophic. Self-driving cars, wearable and implantable medical devices, smart buildings, and critical infrastructure are some high-profile examples that underscore modern CPS’s security and safety concerns. In the last decades, safety concerns have received considerable attention in the design of CPS, while security analysis is left as an afterthought for later stages. This paradigm results in a costly and lengthy development process due to high-security validation costs. We believe that the security considerations should be elevated as primary design drivers and safety ones to tackle the design challenge of modern CPS.

About the Digital Futures Postdoc Fellow
Siyuan Liu is a Postdoctoral researcher at the Division of Decision and Control Systems at KTH. Before joining KTH, she worked as a research assistant at the Institute of Informatics at Ludwig-Maximilian University of Munich (LMU), Germany, from 2019 to 2022. She received her B.E. degree in Automation Science in 2014 and her M.E. in Control Engineering in 2017 from Beihang University, China. She received her PhD in Electrical Engineering from the Technical University of Munich (TUM), Germany, in 2022. Her research interests include safety and security in cyber-physical systems, compositional analysis of large-scale hybrid systems, and automated verification and control of nonlinear control systems.

Main supervisor
Dimos Dimarogonas, KTH

Co-supervisor
Marco Molinari, KTH
Jana Tumova, KTH

Watch the recorded presentation at the Digitalize in Stockholm 2023 event.

About the project

Objective
This project aims to improve a Non-Axiomatic Reasoning System design and combine it with state-of-the-art Deep Learning models for perception. This allows the system to be applied in real-world environments, intending to enhance the autonomy of robots where human intervention is to be kept at a minimum. Application-wise, the system is expected to autonomously perform inspection and maintenance operations of city infrastructure such as power plants. This will ultimately lead to new digitisation technology, which can help solve environmental and societal problems.

Background
The human’s ability to reason has evolved to adapt to difficult situations and changes in the environment faster than current AI models allow. Animals that reason effectively outsmart other species and gain key survival advantages. Non-Axiomatic Reasoning can explain most of these cognitive abilities and provides a roadmap for cognitive enhancements based on psychological and neuroscientific insights. Also, Learning can be explained as inductive reasoning using Non-Axiomatic Logic, an aspect most reasoning systems lack while being a key aspect of intelligence.

About the Digital Futures Postdoc Fellow
Patrick Hammer is a postdoc researcher at Stockholm University, Department of Psychology, working with Robert Johansson and Pawel Herman. Before joining Stockholm University, he got his PhD in Computer Science (AI track) at Temple University, Pennsylvania, United States, where he was a full-time research assistant of Pei Wang. His research interests include Artificial Intelligence, Reasoning Systems, Autonomous Robots, Machine Learning, Deep Learning and Cognitive Science.

Main supervisor
Robert Johansson, Associate Professor at Stockholm University.

Co-supervisor
Pawel Herman, Associate Professor, Computer Science, Division of Computational Science and Technology at KTH.

Watch the recorded presentation at the Digitalize in Stockholm 2023 event.

About the project

Objective
Our research project aims to propose a complete, fully distributed scheme in intelligent task allocation and security-based controller synthesis with relative displacement or bearing-only information for networked multi-agent systems like UAVs or UGVs. The designed methods are expected to be used directly in multiple unmanned autonomous systems. The research plays a significant role in improving the autonomy and swarming of the pursuit system and has great application potential in search and rescue, intelligent transportation, and public protection.

Background
Multiple autonomous unmanned systems, such as unmanned aerial vehicles, unmanned underwater vehicles, and unmanned ground vehicles, are projected to play important roles in many industrial and societal applications, such as search and rescue, cooperative payload carrying, and warehouse management. In the community of multi-agent control, the pursuit-evasion game has been a hot topic whose objectives are to achieve optimal sensor placement intelligent and distributed task allocation, following which the control problem is formulated as a distributed optimization problem with a defined performance index. The objectives of this project thus include the exploration of cooperative and distributed optimal control strategies with relative position measurements or bearing-only information.

About the Digital Futures Postdoc Fellow
Panpan Zhou is a postdoctoral researcher at the Department of Mathematics at KTH Royal Institute of Technology. Before joining KTH, she was a Postdoctoral Fellow in the Department of Mechanical and Automotive Engineering at the Chinese University of Hong Kong. She received her Bachelor’s degree in the School of Automation from Northwestern Polytechnical University, Xi’an, China, in 2017 and her PhD from CUHK in 2021. Her research interests include control theory and applications of multi-agent systems and motion planning of micro aerial vehicles.

Main supervisor
Xiaoming Hu, professor, Department of Mathematics, KTH.

Co-supervisor
Bo Wahlberg, professor, Division of Decision and Control Systems, KTH.

Watch the recorded presentation at the Digitalize in Stockholm 2023 event.

About the project

Objective
The objective of this project is to learn distributed control policies for multi-robot systems that scale on-demand, in multi-laterally evolving complex dynamic environments.

Background
Multi-robot systems such as drone swarms offer unparalleled advantages in assisting ground-based human-robot teams in humanitarian, and disaster-response missions thanks to their ability to operate in remote, communication-denied environments. However, in practical deployments such systems must learn how to balance various auxiliary objectives on-demand for maximizing the collective utility in highly dynamic environments. For example, a robust drone swarm control policy must prioritize the most severely affected area in a disaster-response event, and direct more robots to tackle it, on-the-fly.

In this project, we aim to design a framework that can help us learn such control policies while taking into account the robot’s varying capabilities. Upon designing such a framework, we further understand that we can minimize the human bias in hand-engineered task prioritizing.

About the Digital Futures Postdoc Fellow
Malintha Fernando received his Ph.D in 2023 from Indiana University, Bloomington (USA). His doctoral research focused on designing cooperative scalable control policies for multi-drone systems that are robust to communication failures. Such policies lend themselves to many Urban Air Mobility (UAM) applications such as autonomous parcel delivery, where decentralized decision-making under local information is required to achieve the necessary scalability over the geographical span and in the number of vehicles.

Prior to joining KTH, Malintha worked as a visiting lecturer in Machine Learning at Indiana University. He has completed an internship at Open Robotics, Mountain View, and completed his undergraduate education from University of Moratuwa, Sri Lanka.

Main supervisor
Silun Zhang, Assistant Professor, Division of Optimization and System Theory, Department of Mathematics, KTH.

Co-supervisor
Petter Ögren, Professor, Division of Robotics, Perception and Learning (RPL) at KTH.

About the project

Objective

The overall objective is to develop and evaluate AI-based and classical optimization (mathematical programming) approaches for sensor placement and control, data processing, communication, and motion planning.

This project aims to develop novel algorithms and computational methods for optimal planning, deployment, and operations of a network of AUVs and sensors in undersea environments. These algorithms will focus on optimizing the placement of sensors and movement of AUVs across the region, paying attention to relevant objectives including coverage and communication robustness. They will also focus on processing locally collected AUV measurements (e.g., sonar data) for situational awareness tasks (e.g., the emergence of an adversarial threat). The project will consider classical model-based approaches, AI/ML approaches, and hybrid approaches to developing these algorithms, comparing and contrasting them under different environmental settings and dynamics.

Of particular interest in this project is coordinating sensors and fleets of autonomous underwater vehicles (AUVs) to patrol regions of the ocean. However, these settings pose unique challenges in placement/motion planning such as limited communication, computations, and data processing. The communication challenges are mainly dealt with in a second sub-project led by the researchers at Purdue University. 

Background

Surveillance systems are becoming increasingly reliant on the ability of autonomous networked agents to conduct intelligence, surveillance, and reconnaissance (ISR) tasks. Much recent effort has been devoted to AI/ML-based approaches for augmenting such systems, though pinpointing exactly when AI/ML gives clear-cut advantages over traditional optimization and analytics-based is still an open question. Moreover, while many existing efforts in autonomy are focused on systems of drones, sensors, and other vehicles that operate above the surface, little attention has been paid to undersea ISR settings. The project focuses on the undersea setting, bringing together expertise from different fields, and forms a new line of collaboration between KTH, Purdue University, and Saab.

Crossdisciplinary collaboration

This project is part of a larger collaboration between Saab, KTH, and Purdue University. The project focuses on developing novel algorithms and computational methods for, planning, deploying, controlling, and operating a network of AUVs and different types of sensors over contested undersea environments.

The project brings together expertise from Applied Mathematics, Optimization, Electrical Engineering, and expertise in Underwater Environments.

Participating in the project:

Collaborators in the larger project:

About the project

Objective

The overall goal of the project is to develop optimization frameworks to assist in process design, learning accurate models from process data, and support optimal decision-making. The optimization algorithms and frameworks will be tailored towards the needs and processes of interest for LKAB. The type of optimization problems considered in the project falls into the category of “Gray-box” optimization, where parts of the optimization problem are known analytically, but some parts are given by a simulator. The simulator is not necessarily a complete black box, but the simulation model can be too complex to be integrated directly into the optimization model.

Designing industrial processes is a challenging task and often requires the use of advanced design and simulation software. For example, accurately describing the process and physical phenomena involved often requires the solution of systems of partial differential equations (PDEs) or a Computational Fluid Dynamics (CFD) simulation. Therefore, evaluating the impact of simply changing a single design parameter can require running a CFD or solving a large system of PDEs and may also require access to chemical and thermodynamic libraries. Solving such systems is a challenge on its own, and in practice, requires the use of advanced simulation software. The main downside of such software is the inherent black-box nature; you can evaluate a single specific design choice at a time, but the end-user does not gain more knowledge. In practice, when such software is used for process design the engineers often follow some common rules of thumb and trial and error to come up with a good design that can be evaluated by software and later put into production. However, the resulting design might be far from globally optimal, i.e., there might exist a far more superior design. Employing such a sub-optimal solution can, e.g., result in increased environmental impact due to a higher energy consumption, increased raw material usage and waste. Furthermore, when used in investment planning and feasibility studies, the sub-optimal designs can cause superior technologies/solutions to seem unreasonably expensive and maybe even economically infeasible. Therefore, there is a strong need for frameworks that enable the combination of simulation software with optimization algorithms to find optimal process designs. 

Background

LKAB is an international mining and minerals group that supplies sustainable iron ore, minerals and specialty products. Since 1890, the company has evolved through unique innovations and technology solutions and is driven forward by more than 4,500 employees in 12 countries. The company is the largest supplier of iron ore in the European Union and a key player in the transformation of the iron and steel industry towards sustainability. LKAB’s goal is to develop carbon-free processes and products by 2045.

The work to eliminate carbon emissions creates new challenges and opportunities. Potential routes and new ideas are continuously being investigated and evaluated. As the production processes are complex and experiments often expensive and time-consuming, the feasibility and potential of new alternatives are often investigated numerically through computer simulations. Currently, computer simulations are used in a black-box fashion, i.e., you can evaluate a specific configuration of process parameters, but you don’t gain any more information or knowledge of the best configurations. Therefore, it is of utmost importance to utilize the full potential of the models and to find the best solutions. This project aims to provide a key component for a systematic model-based design approach.

Increasing computational power, increasing amount of data, and overall digitalization are continuously boosting the potential benefits of extensive computational simulations in process design, but how to best utilize simulators is not trivial. To learn and extract knowledge from the process simulations, the project aims to develop deterministic optimization methods and frameworks for determining optimal designs and operations.

Crossdisciplinary collaboration

The project is a collaboration between LKAB and KTH, where we are focusing on developing optimization algorithms suitable for tackling challenging optimization tasks in industrial production processes. The project brings together expert knowledge from Process Systems Engineering, Applied Mathematics, Optimization, and Process Simulation.

Participating in the project:

About the project

Objective
This project aims to develop and evaluate algorithms for dynamic inference offloading in Scania’s fleet, optimizing the trade-off between model accuracy, latency, and resource consumption. The proposed cascaded inference approach ensures that data is first processed by small models at the edge (M1). Depending on inference confidence, decisions are made in real-time on whether to offload tasks to more complex models (M2-M4) with higher accuracy but increased computational costs.

The core objectives include: (1) designing algorithms that improve inference accuracy while minimizing latency and bandwidth usage; (2) providing theoretical guarantees for the proposed methods, particularly in terms of regret minimization; and (3) benchmarking these algorithms using real-world data from Scania’s operational vehicles. The research will contribute to optimizing deep learning model deployment, ensuring scalable and efficient AI integration in industrial applications while maintaining cost-effectiveness and system reliability.

Background
Deep Learning (DL) models have become a standard for data-driven tasks such as classification and predictive analytics due to their high accuracy. However, their computational and memory demands often require cloud-based deployment, which introduces challenges like latency, bandwidth consumption, and security concerns. In response, edge computing has gained traction, enabling inference on resource-constrained Edge Devices (EDs) such as IoT sensors, mobile devices, and autonomous vehicles. While edge deployment reduces communication delays and enhances data privacy, small models often suffer from lower accuracy.

Scania, a global manufacturer of commercial vehicles, faces similar challenges in deploying DL models across its fleet for tasks such as autonomous driving, predictive maintenance, and sustainable operations. There exists a trade-off between accuracy and efficiency when placing models at different computation points. This project explores cascaded inference, where small models operate locally, and only complex cases are offloaded to more powerful computing resources, balancing accuracy with cost efficiency.

Partner Postdocs
This project brings together experts from multiple disciplines to address the challenges of deploying DL models efficiently across Scania’s fleet. Researchers from machine learning, optimization, embedded systems, and automotive engineering will collaborate to develop cascaded inference strategies that optimize accuracy, latency, and resource usage.

Scania’s senior data scientists, Dr. Sophia Zhang Pettersson and Dr. Kuo-Yun Liang, provide real-world insights into vehicle data, predictive maintenance, and cost modelling. Associate Prof. Lei Feng adds knowledge in Bayesian optimization and deep learning techniques for edge computing.

This collaboration ensures that theoretical advancements in machine learning align with practical deployment challenges in commercial vehicles. By integrating perspectives from academia and industry, the project fosters innovation in scalable AI solutions, leading to efficient, adaptive, and cost-effective DL deployment across connected fleets.

Supervisor
KTH researchers, led by Prof. James Gross, contribute expertise in hierarchical inference, algorithm development, and performance guarantees.