Key points are not available for this paper at this time.
Sensors are central elements of modern engineering systems, supporting essential functions in areas such as healthcare, manufacturing, autonomous systems, environmental monitoring, and smart infrastructure. As sensor networks become larger and more integrated into intelligent and networked environments, ensuring their reliability has become increasingly challenging. Faults such as bias, drift, or signal loss can severely degrade data integrity, system performance, and safety. This paper offers a comprehensive review of sensor fault management, going beyond the scope of previous surveys by addressing the full lifecycle, from fault detection and diagnosis to mitigation and compensation. While most existing reviews tend to focus primarily on detection and diagnosis, this work takes a step further by examining a range of advanced mitigation strategies, including hardware redundancy, analytical reconstruction, AI-based approaches, and improved calibration and maintenance practices. The review also provides a structured classification of fault types, investigates their causes and effects across different application areas, and critically evaluates the strengths, assumptions, and limitations of current fault detection and diagnosis (FDD) methods. By bringing together a broad, cross-domain perspective and a detailed analysis of existing methods, this work aims to be a useful reference for researchers and engineers working on reliable, fault-tolerant, and intelligent sensor systems. It also highlights several ongoing challenges, such as separating actual faults from noise, addressing cybersecurity risks, and ensuring that sensor networks can scale and adapt to the demands of complex, safety-critical environments.
Belgacem et al. (Wed,) studied this question.