LAUSR

Target state estimation is a key technology for unmanned underwater vehicles (UUVs) to achieve target tracking, collision avoiding, formation control, and other tasks. Compared with other measurement methods, underwater measurement has lower reliability due to the uncertainty of sonar detection. In this case, the performance of target state estimation depends heavily on the target motion model. However, the dynamics of UUVs are very complex and nonlinear. Although many state estimation methods for nonlinear systems have been proposed, the complex dynamics of UUV and uncertainty in sonar detection remain challenges for underwater target state estimation problems. This article proposes a gated recurrent unit (GRU)-based particle filter (PF) to improve the performance of the target state estimator for UUVs. A deep neural network framework based on GRUs is used to establish the mapping between the previous measurements and the current target states. This neural network learns the dynamics model of UUVs and recognizes the measurement noise. The proposed filter samples from previous measurements of the target UUV, and the fully trained deep neural network predicts the current states of the sampled particles. The proposed method solves the low accuracy and instability caused by motion modeling errors and system nonlinearity. The simulation results show that the proposed GRU-based PF for UUV state estimation has better accuracy and stability than the traditional state estimation methods. All the simulations are done with the noise which is Gaussian noise that can pass through a nonlinear transformation.