Cyber-physical systems (CPS) are an ensemble of physical systems and computers. These systems are quite complex and work by deep collaboration and integration of computation, communication and control technologies. CPS aims to achieve reliable, safe and dynamic cooperation of the cyber domain with the physical system in real time.
The computational component of the CPS monitors and controls the physical system with feedback loops. An essential element in these structures is the management of the physical systems through analysis of various variables, and the use of computational methods to achieve a deeper understanding of the target environment that is being monitored. This provides timely and precise actions based on sound data.
The application of CPS is everywhere, from aeronautics to biomedical industry, from military systems to mundane traffic control. These systems are at the heart of Industry 4.0 and associated intelligent automation. Many concepts that we just take for granted are actually an application CPS and future possibilities are endless.
How it works
CPS is basically an engineering system that is designed to control and operate certain physical processes through computers. Therefore, there are a few components that are common to most CPS systems like sensors, actuators, a control system and they are all connected through a network. A simplified block diagram of a CPS architecture is shown in Figure 1.
Figure 1. Simplified CPS architecture.
Some might confuse wireless sensor networks (WSNs) and internet of things (IoT) with CPS, but it is worth mentioning here that they are not exactly the same. WSNs provide data support for various applications. Their main function is to sense or collect data from various sensor nodes, process this information and then route it. The data collection and signal sensing are not necessarily linked to a particular object. IoT connects smart devices through internet with the overall goal of perceiving data and processing it intelligently.
CPS is a top layer to these technologies. It is a controllable and scalable intersection of computer and physical systems that may rely on data from IoT and WSN or it may have its own localized system. It depends on the specific application of CPS. To put it in simple terms, the architecture of CPS can be separated into three different layers: physical layer, information layer and user layer. The physical layer comprises of the actual physical equipment and all the associated hardware including sensors and actuators. The information layer is primarily responsible for the collection of data from physical layer and then processing it using various computational and control models. The user layer brings in the human component and gives control to humans over various aspects of the system through human-computer interaction. CPS is a closed loop system.
Towards a smart CPS
The growing link between the physical and cyber domain and the creation of more advanced intelligent technologies has led to the evolution of CPS into next-generation "smart" cyber-physical system (sCPS). This is a complex and pervasive software-intensive framework that is able to effectively manage the processes in the physical world and support a wide variety of new applications by integrating a variety of data sources and by using artificial intelligence (AI) techniques. Pervasive intelligence is ushering in an era of smart systems that recognize physical signals and react based on AI, and are independent of the cloud or the internet.
Unlike a traditional CPS where traditional computational and control techniques are the highlight of the system, sCPS aims to develop a pervasively intelligent entity able to use advanced AI for effective monitoring, decision making and controlling of physical systems. Smart, self-managed and self-configured pervasive systems can improve process efficiency across a variety of applications and help in overcoming many major social and environmental concerns.
The key here is to design sCPS in such a way that its components possess a high degree of autonomy and their collaboration is decentralized, scalable and robust. This multidisciplinary field faces a lot of challenges that require rigorous research before its wide-scale implementation becomes feasible and economical. Designing of an intelligent and energy efficient sensing infrastructure is not as easy as it sounds and there are a lot of speed bumps and roadblocks in this journey. Furthermore, the development of a context-aware and self-adaptive system is also not an easy feat.
Data processing and analytics is expected to play a huge role in the future design of sCPS. A huge treasure trove of data is available from social media ranging from room temperatures to geographical locations. Cleaning, processing and extracting useful information from this social-cyber-physical data can play a pivotal role in developing a beneficial sCPS for certain applications. WSNs can also serve as a data sensing and collection infrastructure for a variety of sCPS architectures.
In addition to the data analytics and sensing technologies, energy management is also a crucial component of the sCPS. The physical side of CPS is mostly composed of many battery-powered devices and research into the energy efficiency and management of these devices is critical for the development of economical and sustainable sCPS. Furthermore, reliability and security issues become more prominent as the number of connected devices increases. sCPS must have the ability to adapt to the physical environment, withstanding cyber and physical attacks and maintaining robustness and integrity of data.
The basic components and their associated requirements outlined here present an overall perspective of current and future sCPS. The technology is not yet mature and is continuously evolving. However, this is another thread that will be woven into the fabric of the smart, self-driving and self-sufficient AI based technological future of the engineering systems.
