Research on GRU Neural Network INS/GNSS Fusion Navigation with GNSS Denied
The primary challenge faced by the INS/GNSS is to achieve reliable and cost-effective positioning during GNSS outages.Under certain conditions,such as tunnels,tall buildings in urban areas,and adverse weather conditions,prolonged GNSS signal loss may occur,rendering the INS/GNSS fusion navigation system to degrade to a standalone inertial navigation system.To address this issue,a novel INS/GNSS fusion navigation algorithm based on GRU neural networks in the presence of GNSS denial is proposed.This algorithm operates by utilizing the inertial navigation parameters as inputs to the GRU neural network when GNSS signals are available,while simultaneously utilizing the three-dimensional position in-formation provided by GNSS as the output of the GRU neural network and training it.Subsequently,when GNSS signals disappear,the inertial navigation parameters are used as inputs to the trained GRU model to obtain the three-dimensional position information of GNSS,enabling INS/GNSS fusion navigation even in the presence of GNSS signal denial.The simu-lation results,with a maximum positioning error of 5.646 m,demonstrate that this algorithm effectively ensures the accura-cy and robustness of INS/GNSS fusion navigation under GNSS signal denial.
NETL
NSTL
万方数据
