Embedded Systems, Cyber Physical Systems & Internet of Things
Project overview
Here you will find a number of projects we have realized in the fields of Embedded Systems, Cyber Physical Systems (CPS) & Internet of Things (IoT). Use the ‘+’-sign to see additional information.
Start date: 01.01.2021
End date: 31.12.2025
Funded by: Bayerisches Staatsministerium für Wissenschaft und Kunst (StMWK)
Local head of project:
Local scientists:
Abstract
The Augsburg AI Production Network is an association of the University of Augsburg with the Fraunhofer Institute for Foundry, Composite and Processing Technology IGCV and the Center for Lightweight Production Technology of the German Aerospace Center (DLR). The goal is joint research into AI-based production technologies at the interface between materials, manufacturing technologies and data-based modeling.
The vision of the AI production network is highly modular material-optimized production. In this context, AI technologies are to be addressed along the entire value chain. Artificial intelligence is to play a central role in process optimization and control, the material-appropriate design of products, and the planning of production processes. The AI production network is to research the necessary technologies and help companies quickly implement these approaches in their environment.
Further information about the AI Production Network Augsburg
Start date: 01.10.2020
End date: 30.09.2022
Funded by: Zentrale Innovationsprogramm Mittelstand (ZIM) des Bundesministerium für Wirtschaft und Energie (BMWi)
Local head of project::
Local scientists:
Abstract
Within the scope of CBMD, the University of Augsburg has investigated, among other things, how contracts between different components can be validated during their composition. Furthermore, it was investigated how contracts as interface state machines fit together with the specification of the behavior (state machine) of the component. In the course of the project, it became apparent that contracts also make a significant contribution in the context of security. For example, a check of the contracts at runtime can rule out the possibility of a component being operated "incorrectly". Due to dependencies, however, it can also happen that a function of one component calls a function of another component without the call being made via the interface secured with a contract. Therefore, in this project, a static analysis of the dependencies in the code with integration into the Contracts is to ensure that, firstly, security-relevant functions and data of a component and, secondly, side effects between components can be identified and subsequently checked and analyzed in more detail by tests on the real system.
Start date: 01.07.2019
End date: 30.06.2022
Funded by: Horizon 2020 (H2020) - ECSEL
Local head of project:
Local scientists:
Abstract
In recent years, Cyber Physical Systems (CPS) technologies have become a game changer in strategic sectors such as Automotive, Energy and Industry Automation, where Europe is a world leader. In fact, CPS is a key driver for the innovation capacity of European industries, large and small, generating economic growth and supporting meaningful jobs for citizens.
CPS4EU proposes to address technical issues and organizational issues in an integrated way. Hence, CPS4EU promotes a high level of sharing, so that an operational ecosystem, with adequate skills and expertise all along the value chain can enable, at the end of the project, the European industry to lead strategic markets based on CPS technologies.
The ultimate objective of CPS4EU is to strengthen the CPS value chain by creating world class European SMEs and by providing CPS technologies that in turn will sustain the leadership of the large European groups in key economy sectors and, in this way will stimulate innovative products to support the massive digitization increasingly integrated into our everyday environment.
To achieve these goals CPS4EU will:
- Develop 4 key enabling technologies (computing, connectivity, sensing, cooperative systems)
- Incorporate these CPS modules through pre-integrated architectures and design tools
- Instantiate these architectures in dedicated use cases from strategic application: automotive, smart grid and industry automation
- Improve CPS awareness and usage for all industrial sectors
Start date: 01.04.2018
End date: 31.03.2020
Funded by: Zentrale Innovationsprogramm Mittelstand (ZIM) des Bundesministerium für Wirtschaft und Energie (BMWi)
Local head of project: Prof. Dr. Bernhard Bauer
Local scientists: Reinhard Pröll
Abstract
The aim of this research project is to automate the evaluation of existing tests ("Test the Test", T3) by means of Fault Injection as well as mutations of the test object (system under test) on the software side and on the hardware side to improve the quality of the tests. To this end, existing approaches to software and hardware tests will be supplemented by a quality analysis of test cases in order to meet the ever-increasing security requirements of embedded systems and to adapt the tests semi-automatically to the test results.
In addition to the classical approaches for determining test quality, T3 aims at a better way of evaluating tests. On the one hand, this is to be done by means of so-called "front-loading" of test activities, i. e. tests in early phases of development (design time) and their evaluation. On the other hand, a (semi-)automatic improvement of the test quality is to be achieved by appropriate adaptation and combination of classical code metrics. This evaluation is to be made possible in a similar way across different integration levels. To this end, the results of these developments will be integrated into specific existing software and hardware testing tools of the project partners. The results are evaluated through case studies.
Project start: 01.07.2017
Project end: 30.06.2019
Funded by: Zentrale Innovationsprogramm Mittelstand (ZIM) des Bundesministerium für Wirtschaft und Energie (BMWi)
Local head of project:
Local scientists:
Abstract
Nowadays software functions are not operated in isolation from each other, but usually there are a multitude of dependencies between them. For the manufacturers of individual functions, as well as for the function integrator, it is therefore very difficult to impossible to completely oversee all interactions between the inherent states. This results in complexity effects such as emergence, common mode effects, unwanted activation of operating states, hidden links and dis-functionalities. The aim of the project is therefore to define and implement a hierarchically organized, computer-based development platform for SW-intensive systems that implements the contract-based design paradigm consistently and formally. Accordingly, the platform should be structured hierarchically as well as modularly in order to be able to follow both a top-down (new development) and a bottom-up development process (existing components/subsystems) and to contain all necessary design and test modules that are necessary to carry out the development steps across all process levels in the sense of the CBD paradigm. Evaluation takes place via a case study.
Project start: 01.01.2016
Funded by: Universität Augsburg
Local scientists:
Abstract
Within the Autonomous Driving Lab innovative concepts in the area autonomously driving vehicles are developed and attempted based on vehicles models on a scale of 1:8.
Thereby current challenges of the automotive industry and related research fields are adressed and solutions focusing flexibility and adaptability are emerged.
Project start: 01.10.2016
Project end: 30.09.2019
Funded by: BMBF (Federal Ministry of Education and Research)
Local head of project: Bernhard Bauer
Local scientists:
Christoph Etzel
Christian Saad
Abstract
Development Processes, Tools and Platforms for Safety-Critical Multicore Systems.
Project start: 01.01.2016
Project start: Universität Augsburg
Local scientists:
Christian Saad
Abstract
AUTOSAR ("Automotive Open Systems Architecture") is the de facto standard for automotive ECU software and provides a consistent software architecture as well as uniform description and configurations formats. However, there is a shortage of tools which work directly on AUTOSAR models and do not use proprietary (and often simplified) intermediate models.
Together with the Continental Automotive GmbH, our professorship developed a tool named "AutoAnalyze", which conducts a data-flow analysis on the most fine-granular AUTOSAR level, visualizes the dependencies between the functional blocks, detects potential data consistency conflicts and provides support for resolving them, e.g., by imposing, modifying or removing timing constraints. Hereby, the model is being validated for an execution on single- and multi-core platforms.
Most often, an intended execution on multi-core platforms does not lead to software being re-created from scratch but rather to migrating existing legacy software. Therefore, our tool also supports the required working steps of partitioning (splitting the software into a disjoint set) and mapping (assigning the software parts to cores/execution units) with the help of a previously performed region analysis as well as derived initial solutions from it.
Description
Project start: 01.01.2016
Funded by: Universität Augsburg
Local scientists: Thomas Driessen
Abstract
It is common knowledge in Software Engineering that the earlier in development an error of a system is found, then the lower are the costs for its correction. This is especially the case if the system under development is an embedded or safety critical one, where not only a system’s software, but also its corresponding documentation or hardware is affected by changes.
In this context, Model-Driven Development (MDD) aims to shift most aspects of a system’s software implementation into earlier phases of the development e.g., software design or system design. Therefore, we concentrate in this work on shifting the timing and inter-component communication aspects of a system’s software from the implementation phase to the system design phase of a project.
Our approach uses the Architecture Analysis and Design Language (AADL), which is specifically designed for the specification, analysis, automated integration and code generation of real-time, performance-critical distributed computer systems. AADL offers – among other things – standardized semantics for timing and inter-component communication aspects of software components. In our approach, we utilize these semantics to define a mapping between the AADL and the Real-Time Specification for Java (RTSJ). RTSJ is an extension of standard Java for hard and soft real-time applications. With an implementation of this mapping, we then generate AADL semantic-compliant RTSJ code, which preserves the timing behaviour and intercomponent communication defined in an AADL model. Thus, a system designer is capable of designing and performing analyses regarding communication and timing almost completely during design phase, while resting assured that the implementation will reflect his design choices. Simultaneously, programmers are relieved of the monotonic and repetitive task of writing communication- and timing-related code.
The application of our approach is shown via the implementation of an autopilot for quadrocopters. For this purpose the software of the quadrocopter is modelled in AADL and is then generated by our implementation. The case study shows three advantages of our approach over an implementation without code-generation:
- The speed-up of development by letting the programmer focus on application logic instead of writing recurring code concerned with timing and communication.
- A less error-prone transition from the design of a system to its implementation.
- The possibility of an earlier detection of timing- or communication-related errors in the system.
Our further research is aimed at integrating safety-related aspects e.g., error-propagation, into our existing approach by exploiting Java’s exception mechanisms and RTSJ’s asynchronoustransfer- of-control (ATC) mechanisms.
Further information: MBE for Autonomous Vehicles with Real-Time Java and AADL
Project start: 01.01.2015
Project end: 31.12.2016
Funded by: Zentrale Innovationsprogramm Mittelstand (ZIM) des Bundesministerium für Wirtschaft und Energie (BMWi)
Local head of project: Prof. Dr. Bernhard Bauer
Local scientists:
Christian Saad
Abstract
In ReTeC (Redeuction of Test Complexity) the development of an innovative methodology and tooling is focused, whereby the concepts of Model-Driven Software Development (MDSD) and Model-Based Testing will be strongly connected. The project contributors thereby work on a holistic, model-based and object-oriented approach for development and test of embedded systems. The improvement of automated further use and reuse of development artifacts and the reduction of test complexity itself are the major goals of this project. Based upon entrenched development tooling a integrated and consitent solution is created, which covers the whole testing cycle of a embedded system starting from model-in-the-loop down to hardware-in-the-loop tests.
Project start: 01.09.2014
Project end: 31.08.2016
Funded by: Zentrale Innovationsprogramm Mittelstand (ZIM) des Bundesministerium für Wirtschaft und Energie (BMWi)
Local head of project: Prof. Dr. Bernhard Bauer
Local scientists:
Abstract
Safety critical products have to be developed with respect to electronic or programmable systems in accordance with current generic and product-specic safety standards. The safety integrity level (SIL) which is contained in that standard requires defined metrics with respect to reliability and fail-safe stability of implemented safety functions for products. The objective is to reconcile inceasing demand for safety critical systems and to minimize development effort. Model Driven Software Development has established as essential technology for a quick and efficient system development. Among other things it is already possible to generate 100% of code from models but analyses for determination of the SILs are performed manually and indepent from models. This approach is time-consuming, fault-prone, difficult to trace and there is no reusability. Main objective of this project is a holistic approach for conception, specification, analysis, implementation and testing of a tool chain including methodology which supports fundamentally product development as well as their safety relevant certification of a essential model.
Project start: 01.01.2014
Funded by: Universität Augsburg
Local head of project: Philipp Lohmüller
Abstract
Today, safety-critical systems are used in various domains, including, e.g., the automotive sector. Due to the numerous features that are built into the end products today, however, it can happen that safety-critical concerns such as safety, security or timing are violated. It is the aim of this project to ensure an optimal trade-off in order to achieve a maximum degree of safety. Furthermore, requirements often change these days, which means that even safety-critical components are affected and cannot simply be replaced, as they depend on various other components. Therefore, this project presents a Change Impact Analysis, which determines all such components. In addition, safety-critical products such as an automobile can now be configured using a modular system. For example, there are several million possibilities for a modern compact car. In this context we are talking about (software) product lines. However, not every product line has the same security features. Therefore, components with similar safety features are identified in this project in order to reduce complexity and effort.
Project start: 14.12.2012
Project end: 14.12.2014
Duration: 2 Jahre
Funded by: FuE-Programm "Informations- und Kommunikationstechnik" (IuK Bayern)
Local head of project: Prof. Dr. Bernhard Bauer
Local scientists:
Abstract
WEMUCS is a german acronym that stands for "methods and tools for the iterative development and optimization of software for embedded multicore systems". It is a project supported by the research and development programme "Information and Communication Technique Bavaria". Project partners are the companies Gliwa, Infineon, Lantiq, Lauterbach, sepp.med, Siemens, Timing-Architects, and TWT. Continental acts as associated business. Furthermore the following research facilities participate in handling the four work packages: Fraunhofer ESK, Friedrich-Alexander University of Erlangen (Chair of Programming Systems) and the University of Augsburg (Chair for Software Engineering and Programming Languages).
Project start: 01.10.2012
Project end: 30.09.2014
Funded by: FuE-Programm "Informations- und Kommunikationstechnik"des Freistaates Bayern
Local head of project: Bernhard Bauer
Local scientists:
Benjamin Honke
Thomas Driessen
Abstract
A well known and used development standard for software projects is the V-Modell. Especially in the domain of aerospace this standard is widely accepted and thus is used in most cases as a common base for project development. Hereby, each process discipline like requirements analysis, software design or development, is organized in seperate phases. By extending this standard to the V-Modell XT those processes got standardized and adaptable to different project-dependant situations. Therefore, the V-Modell ist definitely the foundation, but has to be extended, regarding issues like coverage of complexity and related domain-specific processes, standards, methods and tools, by a process framework and an underlying data model. This process framework, as well as the underlying data model are not existent in current approaches and shall be presented by this project as an extended solution, based on the "Layered V-Modell" method.
Project start: 01.07.2012
Funded by: Universität Augsburg
Local head of project:
Abstract
Nowadays innovations in the automotive branch are mainly achieved by software. That's why there are sometimes more than 100 electronic control units (ECUs) integrated within a single car that have to interact with each other via several communication channels. In order to learn about the development of automotive software according to the V-Model, the Automotive Software Engineering Lab provides students access to industry-specific tools and real hardware. Independent workouts address standards, protocols and models as well as their practical application. The Automotive Lab enables the simulation of a vehicle and its environment and illustrates how several ECUs cooperate using a FlexRay bus system and thus provide multiple services within a vehicle. A commercial car simulation software allows to replace certain parts of software by own code, so that it can be verified in a real environment.
Project start: 01.07.2012
Project end: 30.09.2019
Funded by: Universität Augsburg
Local head of project:
Abstract
The system of an aircraft is characterized in particular by its strong decentralization, high safety requirements and mandatory qualification. The Avionic Lab was established to give a practical understanding of the special features of software development in the field of avionics and to provide basic knowledge in this domain. The Avonic Lab takes place in close cooperation with our industrial partners. It consists of a simulator (X-Plane), various embedded computers (Beaglebone Black, Raspberry Pi 2), several workstations equipped with domain-specific tools for development and testing, and a quadrocopter from Erlebrain, on which the software will ultimately run. In our internship, in which the functions of an autopilot are to be implemented/extended, avionics-specific contents such as standards, procedure models, qualification or programming languages are taught in a real test environment. Through direct contact with our partners in the industry, students will also be given practical insights into the avionics industry from past, present and future projects, thus enabling an easy entry into the industry.
Translated with www.DeepL.com/Translator (free version)
Validation-driven design for component-based architectures
Project start: 01.06.2009
Project end: 31.05.2012
Funded by: BMBF (Bundesministerium für Bildung und Forschung)
Local head of project: Prof. Dr. Bernhard Bauer
Local scientists:
Dipl.-Inf. (FH) Dipl.-Math. Stefan Fenn
Dipl.-Inf. Christian Saad
External scientists / cooperations:
Verde Consortium
Abstract
The growing complexity of software intensive, real-time embedded systems combined with constant quality and time-tomarket constraints creates new challenges for engineering practices. These systems are developed according to a traditional application of the verification-and-validation cycle, where V&V activities start only when implementation and integration is completed. Many major issues, often related to the architecture and introduced early in the process, are not found until integration and validation. At this point, they are more difficult and more expensive to fix. While preserving the V&V cycle, VERDE is promoting a more iterative and incremental approach to software development that will be driven by the early V&V activities. The two principal goals are to: 1. Develop a solution for iterative, incremental development and validation of RTES that integrates testing and analysis tools; 2. Foster the industrialisation of this solution through a close collaboration between technology providers and end users from different domains, specifically software radio, aerospace, railway and automotive. This ITEA 2 project will be an opportunity for a close collaboration between mature technology providers and end users from different sectors of the Industry, with the overall objective of investigating new directions for the next generation of engineering tools.
Project start: 01.01.2009
Funded by: Universität Augsburg
Local head of project: Prof. Dr. Bernhard Bauer
Local scientists: Christian Saad
Abstract
The goal of the "Model Analysis Framework" is to provide a core framework (based on the Eclipse Modeling Frameworks) along with a development environment that integrates into Eclipse, allowing the implementation of dynamic model analysis.
Data-flow analysis, which is a technique commonly used in compiler construction, serves as a basis for describing and evaluating the dynamic behavior of models. To accomplish this, DFA algorithms will be adapted so that meta model elements can be annotated with data-flow equations in the form of semantic attributes whose calculation rules are specified using OCL.
Use cases include but are not limited to the computation of the cycle time of business processes, the generation of test cases or the analysis of model metrics.