| 1. |
- Bennulf, Mattias, 1992-
(författare)
-
A Control Framework for Industrial Plug & Produce
- 2023
-
Doktorsavhandling (övrigt vetenskapligt/konstnärligt)abstract
- Kundanpassade produkter och korta produktionsserier blir alltmer populärt. Detta har lett till problem för dedikerade tillverkningssystem som är designade för massproduktion. Det krävs ofta långa produktionsserier för att det ska bli en rimlig investering att ställa om produktionen. Därför används människor för tillverkningsuppgifter som ofta ställs om. Denna avhandling fokuserar på konceptet Plug & Produce, som gör det enklare att flytta, lägga till och ta bort resurser från ett tillverkningssystem. Tanken är att resurser placeras i processmoduler som alla har samma fysiska gränssnitt för att kopplas in i tillverkningssystemet. Styrningen av tillverkningssystemet görs av ett multiagentsystem där varje detalj som ska produceras för produkter får en egen agent som representerar detaljen och agerar som styrningsmjukvara. Varje detaljs agent tar hand on sina egna tillverkningsmål genom att kommunicera med resursagenter i systemet som används för styrning av resurserna. I detta arbete, presenteras ett ramverk för Plug & Produce som består av ett konfigurerbart multiagentsystem, samt ett konfigurationsverktyg som kan användas för att definiera agenterna. Arbetet inkluderar metoder för att identifiera inkopplade resurser, kommunikation mellan agenter, schemaläggning som kan undvika konflikter mellan agenter, samt metoder för att automatiskt hitta vägar för transport genom tillverkningssystemet.
|
|
| 2. |
- Al-Tashi, Mohammed, et al.
(författare)
-
Classroom-ready open-source educational exoskeleton for biomedical and control engineering
- 2024
-
Ingår i: Automatisierungstechnik. - 0178-2312. ; 72:5, s. 460-475
-
Tidskriftsartikel (refereegranskat)abstract
- In recent years, robotic arm exoskeletons have emerged as promising tools, finding widespread application in the rehabilitation of neurological disorders and as assistive devices for everyday activities, even alleviating the physical strain on labor-intensive tasks. Despite the growing prominence of exoskeletons in everyday life, a notable knowledge gap exists in the availability of open-source platforms for classroom-ready usage in educational settings. To address this deficiency, we introduce an open-source educational exoskeleton platform aimed at Science, Technology, Engineering, and Mathematics (STEM) education. This platform represents an enhancement of the commercial EduExo Pro by AUXIVO, tailored to serve as an educational resource for control engineering and biomedical engineering courses.
|
|
| 3. |
- Bennulf, Mattias, 1992-
(författare)
-
A User-Friendly Approach for Applying Multi-Agent Technology in Plug & Produce Systems
- 2020
-
Licentiatavhandling (övrigt vetenskapligt/konstnärligt)abstract
- This thesis presents methods for simplifying the use of multi-agent systems in Plug & Produce. The demand for customized products and low volume production is constantly increasing. The industry has for many years used dedicated manufacturing systems where it is difficult and expensive to adapt to new product designs. Instead, factories are forced to use human workers for certain tasks that demand high flexibility and rapid adaption for new product designs. Several solutions have been proposed over the years to create highly flexible automation systems that automatically handles rapid adaption for new products. A concept called Plug & Produce aims at creating a system where resources and parts can be added in minutes rather than days in dedicated systems. One promising solution for implementing Plug & Produce is the distributed approach called multi-agent systems, where each resource and part get its own controller that communicates with each other to reach manufacturing goals. The idea is that the system automatically handles the adaption for new products. However, still today the use of such systems is extremely limited in the industry. One reason is the lack of mature multi-agent systems that are easy to use and that hides the complexity of the underlying agent system from the users. This is a huge problem since these systems tend to be more complex than traditional approaches. Thus, this thesis focuses on simplifying the use of multi-agent systems by proposing various methods for bringing the multi-agent technology for Plug & Produce closer to the industry.
|
|
| 4. |
- Bennulf, Mattias, 1992-, et al.
(författare)
-
Goal-Oriented Process Plans in a Multiagent System for Plug & Produce
- 2021
-
Ingår i: IEEE Transactions on Industrial Informatics. - : Institute of Electrical and Electronics Engineers (IEEE). - 1941-0050 .- 1551-3203. ; 17:4, s. 2411-2421
-
Tidskriftsartikel (refereegranskat)abstract
- This article presents a framework for Plug & Produce that makes it possible to use configurations rather than programming to adapt a manufacturing system for new resources and parts. This is solved by defining skills on resources, and goals for parts. To reach these goals, process plans are defined with a sequence of skills to be utilized without specifying specific resources. This makes it possible to separate the physical world from the process plans. When a process plan requires a skill, e.g., grip with a gripper resource, then that skill may require further skills, e.g., move with a robot resource. This creates a tree of connected resources that are not defined in the process plan. Physical and logical compatibility between resources in this tree is checked by comparing several parameters defined on the resources and the part. This article presents an algorithm together with a multiagent system framework that handles the search and matching required for selecting the correct resources.
|
|
| 5. |
- Cronrath, Constantin, 1990, et al.
(författare)
-
Formal Properties of the Digital Twin-Implications for Learning, Optimization, and Control
- 2020
-
Ingår i: IEEE International Conference on Automation Science and Engineering. - 2161-8070 .- 2161-8089. ; 2020-August, s. 679-684
-
Konferensbidrag (refereegranskat)abstract
- Digital twins are regarded as enablers of smart and autonomous manufacturing systems. The digital twin concept essentially refers to a ultra-realistic digital model of a products or system, coupled by a bidirectional automated data exchange, used for simulation, optimization, and control. Although the concept has gained significant attention, its conceptual basis is still weak. We review common definitions and descriptions of digital twins and refine the concept in system theoretic terms. With this sharpened perspective on digital twins, we sketch out three basic problems that need to be solved to turn a digital model into a digital twin. To that end, we call attention to challenges that need to be researched into.
|
|
| 6. |
- Cronrath, Constantin, 1990, et al.
(författare)
-
How Useful is Learning in Mitigating Mismatch Between Digital Twins and Physical Systems?
- 2024
-
Ingår i: IEEE Transactions on Automation Science and Engineering. - 1558-3783 .- 1545-5955. ; 21:1, s. 758-770
-
Tidskriftsartikel (refereegranskat)abstract
- In the control of complex systems, we observe two diametrical trends: model-based control derived from digital twins, and model-free control through AI. There are also attempts to bridge the gap between the two by incorporating learning-based AI algorithms into digital twins to mitigate mismatches between the digital twin model and the physical system. One of the most straightforward approaches to this is direct input adaptation. In this paper, we ask whether it is useful to employ a generic learning algorithm in such a setting, and our conclusion is "not very". We denote an algorithm to be more useful than another algorithm based on three aspects: 1) it requires fewer data samples to reach a desired minimal performance, 2) it achieves better performance for a reasonable number of data samples, and 3) it accumulates less regret. In our evaluation, we randomly sample problems from an industrially relevant geometry assurance context and measure the aforementioned performance indicators of 16 different algorithms. Our conclusion is that blackbox optimization algorithms, designed to leverage specific properties of the problem, generally perform better than generic learning algorithms, once again finding that "there is no free lunch".
|
|
| 7. |
- Cronrath, Constantin, 1990, et al.
(författare)
-
Relevant Safety Falsification by Automata Constrained Reinforcement Learning
- 2022
-
Ingår i: IEEE International Conference on Automation Science and Engineering. - 2161-8070 .- 2161-8089. ; 2022-August, s. 2273-2280
-
Konferensbidrag (refereegranskat)abstract
- Complex safety-critical cyber-physical systems, such as autonomous cars or collaborative robots, are becoming increasingly common. Simulation-based falsification is a testing method for uncovering safety hazards of such systems already in the design phase. Conventionally, the falsification method takes the form of a static optimization. Recently, dynamic optimization methods such as reinforcement learning have gained interest for their ability to uncover harder-to-find safety hazards. However, these methods may converge to risk-maximising, but irrelevant behaviors. This paper proposes a principled formulation and solution of the falsification problem by automata constrained reinforcement learning, in which rewards for relevant behavior are tuned via Lagrangian relaxation. The challenges and proposed methods are demonstrated in a use-case example from the domain of industrial human-robot collaboration, where falsification is used to identify hazardous human worker behaviors that result in human-robot collisions. Compared to random sampling and conventional approximate Q-learning, we show that the proposed method generates equally hazardous, but at the same time more relevant testing conditions that expose safety flaws.
|
|
| 8. |
- Gleeson, Daniel, 1988, et al.
(författare)
-
Generating Optimized Trajectories for Robotic Spray Painting
- 2022
-
Ingår i: IEEE Transactions on Automation Science and Engineering. - 1558-3783 .- 1545-5955. ; 19:3, s. 1380-1391
-
Tidskriftsartikel (refereegranskat)abstract
- In the manufacturing industry, spray painting is often an important part of the manufacturing process. Especially in the automotive industry, the perceived quality of the final product is closely linked to the exactness and smoothness of the painting process. For complex products or low batch size production, manual spray painting is often used. But in large scale production with a high degree of automation, the painting is usually performed by industrial robots. There is a need to improve and simplify the generation of robot trajectories used in industrial paint booths. A novel method for spray paint optimization is presented, which can be used to smooth out a generated initial trajectory and minimize paint thickness deviations from a target thickness. The smoothed out trajectory is found by solving, using an interior point solver, a continuous non-linear optimization problem. A two-dimensional reference function of the applied paint thickness is selected by fitting a spline function to experimental data. This applicator footprint profile is then projected to the geometry and used as a paint deposition model. After generating an initial trajectory, the position and duration of each trajectory segment are used as optimization variables. The primary goal of the optimization is to obtain a paint applicator trajectory, which would closely match a target paint thickness when executed. The algorithm has been shown to produce satisfactory results on both a simple 2-dimensional test example, and a non-trivial industrial case of painting a tractor fender. The resulting trajectory is also proven feasible to be executed by an industrial robot.
|
|
| 9. |
- Gleeson, Daniel, 1988, et al.
(författare)
-
Robot spray painting trajectory optimization
- 2020
-
Ingår i: IEEE International Conference on Automation Science and Engineering. - 2161-8070 .- 2161-8089. ; 2020-August, s. 1135-1140
-
Konferensbidrag (refereegranskat)abstract
- In the manufacturing industry, spray painting is often an important part of the process. Especially in the automotive industry, the perceived quality of the final product is closely linked to the exactness and smoothness of the painting process. For complex products or low batch size production, manual spray painting is often used. But in large scale production with a high degree of automation, the painting is normally performed by industrial robots. There is a need to improve and simplify the generation of robot trajectories used in industrial paint booths. A method for spray paint optimization is presented, which can be used to smooth out an initial trajectory and minimize paint thickness deviations from a target thickness. By fitting a spline function to experimental data, an applicator footprint profile is determined, which is a two-dimensional reference function of the applied paint thickness. This footprint profile is then projected to the geometry and used as a deposition model at each point along the trajectory. The positions and durations of all trajectory segments are used as optimization variables. They are modified with the primary goal to obtain a paint applicator trajectory, which will closely match a target paint thickness when executed. The algorithm is shown to produce satisfactory results on both a simple 2-dimensional test example, and a nontrivial industrial case of painting a tractor render. The final trajectory shows an overall thickness close to the target thickness, and the resulting trajectory is feasible to execute directly on an industrial robot.
|
|
| 10. |
- Hagebring, Fredrik, 1985, et al.
(författare)
-
On Optimization of Automation Systems : Integrating Modular Learning and Optimization
- 2022
-
Ingår i: IEEE Transactions on Automation Science and Engineering. - : Institute of Electrical and Electronics Engineers Inc.. - 1545-5955 .- 1558-3783. ; 19:3, s. 1662-1674
-
Tidskriftsartikel (refereegranskat)abstract
- Compositional Optimization (CompOpt) was recently proposed for optimization of discrete-event systems of systems. A modular optimization model allows CompOpt to divide the optimization into separate sub-problems, mitigating the state space explosion problem. This paper presents the Modular Optimization Learner (MOL), a method that interacts with a simulation of a system to automatically learn these modular optimization models. MOL uses a modular learning that takes as input a hypothesis structure of the system and uses the provided structural information to split the acquired learning into a set of modules, and to prune parts of the search space. Experiments show that modular learning reduces the state space by many orders of magnitude compared to a monolithic learning, which enables learning of much larger systems. Furthermore, an integrated greedy search heuristic allows MOL to remove many sub-optimal paths in the individual modules, speeding up the subsequent optimization.
|
|
