Design and implementation a novel inertial sensor errors model in extended Kalman filter based on genetic algorithm

In this paper, a novel algorithm for the real-time modeling of the inertial sensor errors is developed based on the genetic algorithm. In the inertial navigation systems, failure to compensate inertial sensor errors not only could result in exponentially increasing of positioning errors, but also leads to loss of precision in the estimation of other states. This subject indicates the importance of providing an accurate algorithm for sensor errors modeling in the Inertial Navigation Systems (INS). In this respect, a genetic algorithm is proposed in this study for modeling of inertial sensor errors. The main aim of designing this algorithm is to improve the estimation accuracy of position, velocity and attitude of the INS in the presence of uncertainties in inertial sensors. Some vehicular tests have been carried out in order to comparison, implementation and verification of the proposed integration scheme. The experimental results indicate that the proposed algorithm enhanced the estimation accuracy compared with the conventional Gauss-Markov model. Therefore, in general, the proposed method will improve the results of inertial navigation.