A collaborative approach to fault tolerance in cyber-physical systems

Abstract

The reliability and robustness of cyber-physical systems are critical aspects of the current industrial landscape. The high level of autonomous and distributed components associated with many devices makes cyber-physical systems prone to faults. With this in mind and contextualized with the current era of Industry 4.0, fault tolerance is essential for maintaining the robustness and resilience of systems facing unforeseen or undesirable disturbances. Despite their importance and benefits, traditional approaches for fault tolerance, such as local or monolithic centralized, do not correspond with the accelerated technological evolution pace over the past two decades. Driven by the advent of digital technologies such as the Internet of Things, cloud and edge computing, and artificial intelligence associated with enhanced computational processing and communication capabilities, the traditional fault tolerance approaches are out of sync with contemporary and future systems. Consequently, these approaches are limited in achieving the maximum benefits enabled by integrating these technologies, such as robustness related to fault tolerance accuracy and performance improvements, maintaining the flexibility and scalability fundamental for cyberphysical systems operations. Accordingly, this thesis proposes a collaborative fault tolerance approach for cyber-physical systems called Collaborative Fault-Star (CF*), a unified term that refers to the collaboration on the detection, diagnosis, and recovery stages of fault tolerance. The proposed approach takes advantage of cyber-physical components’ inherent data analysis and communication capabilities and is based on multi-agent system principles, where components are self-fault tolerant and adopt collaborative and distributed intelligence behavior when necessary to transcend their individual fault tolerance capabilities. This thesis introduces the backbone and infrastructure of CF*, highlighting the components, agent behavior, functionalities, and interaction protocols to explore the benefits of communication and collaboration between agents to enhance fault tolerance. The CF* is a generic approach, and can be instantiated for diverse cyber-physical system domains as well as for different applications. Experiments were conducted on a developed laboratory-scale warehouse, focusing on the fault detection and diagnosis stage for temperature and humidity sensors in the warehouse racks. The experimental results confirmed the accuracy and performance improvements under CF* compared with the local approach and competitiveness compared to a centralized approach.