An Improved Recursive Interacting Multi-Model Algorithm Based on Genetic Algorithm

Abstract In traditional multi-model algorithm, the models' parameters are set in advance without any changes in the following calculating steps, sometimes these parameters areinitialized randomly. In order to set the parameters more accurate, algorithm based on genetic algorithm (GA) is proposed. Combined with recursive interacting multi-model (RIMM) algorithm proposed by Leigh A. Johnston, the optimal searching ability of GA is used to identify the best model parameters in one searching period. The proposed method is applied to inertial navigation system (INS)/celestial navigation system(CNS)/GPS integrated navigation for missile. The experiment results exhibit satisfactory precision and searching velocity. By determining the models' parameters of the whole filtering adaptively, it reaches a balance between two incompatible goals: “rapid velocity” and “high precision”.