EXNAV: an intelligent sensor integrator

The authors describe the architecture and design of the EXNAV intelligent system that is under development for integrating various navigation sensors with different performance characteristics for an optimal navigation solution. EXNAV, which mixes reasoning with conventional processing for problem solving, consists of an executive and a set of cooperating mini-experts, each responsible for a particular sensor. Using classical Kalman filter algorithm, EXNAV combines all sensor information based on the best sensor configuration. In addition, it provides facilities for sensor-level diagnosis, reconfiguration, and monitoring. Each mini-expert has specialized knowledge about the diagnostics and procedures associated with a particular sensor. The knowledge engineering approach involved in the EXNAV design is discussed