Advancing smart, autonomous cyber-physical systems A new suite of design methods and tools is putting an Italian company at the forefront of the move towards the next generation of autonomous cyber-physical systems. Digital Economy Security © Links Foundation Cyber-physical systems (CPS) sit at the crossroads of the cyber and physical worlds. From autonomous vehicles to smart grid systems, CPS are behind many of the large-scale, safety-critical technologies that will define our future. As embedded systems with communication capabilities, CPS are leveraged to monitor and control the physical world and its associated processes. They create feedback loops where physical processes and computations affect each other. Yet despite this potential, CPS development remains limited, with several challenges that must be addressed. “While the increase in CPS adoption has resulted in a general maturation in available solutions for CPS development, we lack a consistent, consolidated approach for their integration,” says Claudio Pastrone, head of IoT and Pervasive Technologies Research at the LINKS Foundation, an Italian research and innovation centre, and coordinator of the EU-funded CPSwarm project. “As a result, CPS development remains a complex and error-prone task, often requiring the use of many different tools.” To address this challenge, the CPSwarm project has simplified the process of integrating CPS into ‘swarms’. CPS swarms are complex herds of heterogeneous CPS that collaborate based on local policies. They are collectively capable of solving complex, industrial-driven, real-world problems that possibly involve human interaction. Model-centric design and predictive engineering CPSwarm’s goal was to develop the tools and approaches that will pave the way towards well-established, model-based and predictive engineering design methodologies and toolchains for the next generation of CPSs. To do this, researchers built a suite of solutions, including artificial intelligence and Internet of Things (IoT) techniques, that ease the development of autonomous and collaborative CPS. Communication, security and safety aspects were also considered. “Model-centric design is at the core of the CPSwarm project, which aimed to create a library of reusable models for CPS design,” explains Pastrone. “Predictive engineering is another pillar of the project. This enables model verification and simulation of collaborative, autonomous CPS behaviour against real-world and hard-to-handle physical data.” A design and simulation environment Out of this work came the CPSwarm Workbench, a fully fledged design and simulation environment to support the iterative, computer-aided, model-based design of CPS. The Workbench contains a set of tools and solutions, including a modelling tool that allows the user to model CPS swarm members and behaviour, as well as environment characteristics. Other key solutions include a modelling library, simulation and optimisation environment, code generator and deployment tool. According to Pastrone, with these tools, a designer can set up collaborative and autonomous CPSs, test the swarm performance in respect of the design goal, and massively deploy solutions towards reconfigurable CPS devices. “The CPSwarm Workbench represents a step change in large-scale CPS engineering, with an expected significant reduction in development time and costs,” he adds. Real-world tested During the project, the CPSwarm solutions were tested in three real-world use cases. “These tests clearly demonstrate that our proposed model-based methodology can help design and develop a CPS swarm application,” concludes Pastrone. “Using the CPSwarm Workbench drastically reduces the time needed to develop an application based on swarms of CPSs.” The project has released the Workbench as open source, and it is freely available to the CPS and robotics communities. Keywords CPSwarm, cyber-physical systems, CPS, autonomous vehicles, smart grid, IoT, Internet of Things, CPS swarms, predictive engineering
