A cooperative expert system architecture for embedded avionics

An underlying requirement for artificial intelligence software in avionics environments is real-time throughput and behavior. By using multiple processors and efficient communication mechanisms, the concept of cooperative expert systems can be used effectively in real-time applications. A hardware and software environment for supporting cooperative expert systems with multiple streamlined set (RISC-influenced) processors and a real-time shared memory multicomputer operating system executive is presented and analyzed. By using the parallelism of cooperative expert systems and RISC-influenced real-time shells in conjunction with conventional preemptive multitasking, the execution speed of artificial intelligence programs can be increased to meet real-time requirements while also maintaining the predictable response characteristics expected of avionics systems