Research on Attitude Solving Algorithm of Towing Cable Based on Convolutional Neural Network Fusion Extended Kalman Filter

Abstract

Marine seismic exploration is an important part of offshore oil and gas exploration, which requires accurate attitude information of submarine towing equipment. Conventional attitude solution algorithm or Kalman filter algorithm cannot satisfy the current requirements of high accuracy, high reliability, strong environmental adaptability and low cost. In view of the low accuracy and poor environmental adaptability of the traditional Kalman filter algorithm, this paper proposes a CNN-EKF fusion attitude calculation algorithm based on the study of the extended Kalman filter (EKF) model and the convolutional neural network (CNN) model. The system noise variance matrix (Q) and the observation noise variance matrix(R)of EKF were optimized by CNN, and the final solution results were obtained. Compared the traditional Kalman filtering model with the CNN-EKF fusion filtering model, experimental results shows that the algorithm improves the accuracy of attitude calculation and enhances the adaptive ability to the environment.