CLIPS-based Expert System Integration into a POSIX-Compliant Real-Time Operating System

Modern embedded systems, such as industrial controllers, medical devices, or autonomous robots, require the integration of artificial intelligence algorithms with guaranteed response times. Expert systems, capable of processing externally received information in the form of facts through predefined rules and making decisions based on them, are becoming critical components of such systems. However, most expert systems are designed for general-purpose environments and are not optimized for operation in real-time operating systems (RTOS). The CLIPS (C Language Integrated Production System) stands out among its counterparts due to its minimalist architecture, open-source nature, and capability for deep integration with the C language. These features make CLIPS a preferred choice for RTOS, where predictability and efficient resource management are critical. This article describes the methodology proposed by the authors for integrating CLIPS into POSIX-compatible RTOS, addressing key compatibility challenges: Dynamic memory management and fragmentation. Single-threaded Rete-based engine. Unpredictable rule execution time. Lack of integration with the RTOS task scheduler.