An anti-interference MIμGPS vehicle integrated navigation algorithm based on IDNN-EKF

Since the MIμGPS has advantages of low-cost and small-size, it can be widely used in the field of vehicle navigation. For the traditional MIμGPS integrated navigation system, the Kalman filter is used to fuse the information of MIMU and GPS to achieve the goal of navigation and positioning of vehicle. GPS provides users with highly accurate three-dimensional position and velocity informations through the Kalman filter correction to obtain accurate navigation results. However, in actual vehicle navigation applications, it is impossible to obtain accurate system model and noise model, which leads the estimation error accumulation and filter divergence sometimes. In addition, there will be varying degrees of GPS outages phenomenon, when the vehicle is driving in different environments. So in this situation, the Kalman filter will not be able to estimate the navigation information accurately, or eventually led to a large error. This paper proposes an anti-interference MIμGPS vehicle integrated navigation algorithm based on IDNN-EKF. On the basis of the Extended Kalman Filter (EKF), the input-delay neural network (IDNN) is added to assist the navigation system and the constraint equations according to the driving characteristics of the vehicle are established to restrain the input-delay neural network during GPS outages. Moreover, an inspecting method of GPS signal quality based on fault detection is also proposed in this paper to inspect the GPS outages. Finally, experimental road tests involving a vehicle navigation system are performed to validate the effectiveness and availability of the proposed method, compared with traditional methods.