A theoretical framework is proposed for using a rule-based system model to modify control operations using intermittent sensor measurements. A set of general rules and a control modification operator are presented. The rules are used to make decisions based on the system condition by simply adjusting the control gains, or perhaps by restructuring a Kalman estimator. Preliminary theoretical results in reconfiguration of Kalman estimator equations are shown and have been simulated with image tracking systems. A prototype intelligent controller using the proposed structure and implemented in Prolog and C is demonstrated for control of a second-order system with simple fusion of two nonindependent sensors.