Testing Cyber-Physical Systems under Uncertainty
Jan. 01, 2015 to Dec. 31, 2017
Cyber-physical Systems (CPSs) link the digital and the physical world. As such, they connect a number of components (e.g. sensors, actuators, hardware and software) to a networked unity. CPSs are increasingly able to provide intelligent services and are associated with a high degree of automation. These intelligent, networked cyber-physical are increasingly used in safety or mission critical domains such as healthcare, automotive, production, energy, and transport.
Cyber-physical systems are used in almost every area of our lives. Therefore a high level of quality as well as reliability, error robustness, safety and security is required. To prevent endangering users or the environment, these properties are particularly important in dynamically changing environments. Such environments are difficult to predict and even more difficult to describe. This degree of unpredictability about the future purpose and place of use or even of future collaboration partners of a networked system is called “uncertainty”. Particularly in safety-critical areas, a risk-oriented handling of uncertainties is important.
Project goals of U-Test
U-Test aims to improve the reliability of cyber-physical systems by provoking, analyzing and evaluating uncertainties that can occur in the systems or their environment. For this purpose, a model-based, search-based test approach is chosen. This achieves a high degree of automation, thus enabling cost savings in the quality assurance of such systems. In detail, U-Test will:
- provide a comprehensive and extensible taxonomy of uncertainties, classifying uncertainties, their properties, and their relationships.
- create an Uncertainty Modeling Framework (UMF) based on existing modeling and testing standards. Using this UMF, uncertainties at various architectural levels as well as a test evaluation function for critical system functionality can be modeled.
- develop an intelligent, search-based method to uncover previously unknown uncertainties. This method is based on a functional model of the system under test as well as uncertainty models in the system or its environment.
- generate test cases that are automated and highly scalable based on the reaction of the system under test. These test cases provoke previously unknown uncertainties in the system and detect impairments to the critical system functionality.
The feasibility and applicability of the methods and concepts developed in the U-Test project will be demonstrated in two industrial case studies: a Geo Sports Case Study (athlete health monitoring during training and match) and Handling Systems Case Study (warehouse logistics).
Contribution from SQC
The System Quality Center (SQC) adapts its experiences in developing advanced modelling and testing methods for safety-critical systems to uncover unknown uncertainties of cyber-physical systems. In addition, technologies developed by SQC are used. This includes:
- ModelBus®: An open source framework for tool integration in software and systems development that automates the execution of complex and error-prone development tasks.
- Focus! MBT: A test modeling methodology and test automation platform focused on automation of dynamic test processes – including automated test design, automated execution as well as automated test evaluation.
U-Test is funded by the European Union's funding program Horizon 2020.