Irregular situations in real-world intelligent systems

Real-world is full of uncertainty. This uncertainty introduces examples of irregular situations (situations that are contrary to the normal routine). Artificial intelligence (AI) has greatly benefited society through automation and intelligent systems. However, real-world situations often involve elements of unpredictability and irregularity, which can pose challenges for AI systems. To address these challenges, researchers have developed various techniques to improve the robustness and adaptability of AI systems. These include methods for handling uncertainty, data pre-processing, explainable AI, safety-critical AI, etc. However, despite these efforts, many open questions and challenges must be addressed to make AI systems more robust and adaptable to real-world situations. In this work, we have defined the possible irregular situations (IS) and introduced the potential solutions to countermeasure such situations. Here, we have surveyed the IS in image, audio, olfactory, and motion intelligence. Further, we have investigated a few of the way-outs and solutions. In addition, we have demonstrated the IS in automated driving depending upon the level of autonomy in autonomous vehicles (AVs) and discussed the safety and privacy issues with a consideration of the safety of the intended functionality (SOTIF) standard. These findings will undoubtedly facilitate research in the direction of future mobility.