A model for probabilistic monitoring and proactive restart of real-time operating systems under intensive state changes in cyber-physical systems

The article presents a model of probabilistic monitoring and proactive restart of real-time operating systems (RTOS) in conditions of intensive changes in the states of cyber-physical systems. The proposed model uses a probabilistic approach to assess the risk of failures, which allows for timely prediction of potential problems and prevention of their critical impact on the functioning of the system. It provides proactive local recovery of problematic components using a dual watchdog mechanism, which includes both hardware and software control levels. This approach allows not only to detect failures at an early stage, but also to promptly respond to deviations from the normal operating mode of the system. A key aspect of the developed model is the ability to locally restart individual tasks before the hardware watchdog timer is triggered. This significantly reduces the overall computational load on the microcontroller, preventing excessive use of resources and reducing the likelihood of a complete emergency restart of the system. Integration of the proposed approach with real-time operating systems, such as FreeRTOS, provides ease of implementation and increases the efficiency of resource management in systems with limited capabilities. In addition, the model allows you to adapt to changing operating conditions and provides flexible tuning of monitoring parameters in accordance with the specifics of the operation of cyber-physical systems. Experimental studies have demonstrated the high effectiveness of the proposed method in reducing downtime and increasing the overall reliability of real-time operating systems. The proposed approach allows you to minimize the consequences of unexpected failures, increase the system's resistance to changes in the external environment and ensure the stability of its operation even in critical operating conditions.