Improved LSTM Neural Network-Assisted Combined Vehicle-Mounted GNSS/SINS Navigation and Positioning Algorithm

Abstract

Aiming at the problem of the combined navigation system of on-board GNSS (global navigation satellite system)/SINS (strapdown inertial navigation system), the accuracy of the combined navigation system decreases due to the dispersion of the SINS over time and under the condition of No GNSS signals. An improved LSTM (long short-term memory) neural network in No GNSS signal conditions is proposed to assist the combination of navigation data and the positioning algorithm. When the GNSS signal is normal input, the current on-board combination of the navigation module’s output sensor data information is used for training to improve the LSTM algorithm and to establish the incremental output of the GNSS position of the mapping of the different weights. In No GNSS signal conditions, using the improved LSTM algorithm can improve the combination of navigation and positioning algorithms. Under No GNSS signal conditions, the improved LSTM training model is used to predict the dynamics of SINS information component data. Under No GNSS signal conditions, the combined navigation filtering design is completed, and the error correction of SINS navigation and positioning information is carried out to obtain a more accurate combination of navigation and positioning system accuracy. It can be seen through the actual test experiment using a sports car in the two trajectories under the conditions of No GNSS signals that the proposed algorithm can be compared with the LSTM algorithm. In testing road sections, the proposed algorithm, when compared with the LSTM algorithm to obtain the northward position that the mean square errors were improved by 55.63% and 76.64%, and the eastward position mean square errors were improved by 43.42% and 54.67%. In a straight-line trajectory, improving the effect’s navigation and positioning accuracy and reliability is significant.

A tightly coupled GNSS RTK/IMU integration with GA-BP neural network for challenging urban navigation

“…The used navigation parameters, such as specific force increments and angular velocity increments extracted from the raw observations of IMU, and the corresponding outputs are GNSS position increments [11][12][13][14] or state vector of the integrated navigation system, that is position and velocity error between GNSS/IMU fusion estimation and IMU estimation [3,[15][16][17][18][19][20]. Yao et al proposed a scheme for GNSS/IMU to represent the vehicle dynamic variations with the current and past 1-step specific force, angular rate and velocity to estimate the GPS position, further enhancing performance of the positioning system during GPS downtime [12].…”

“…Chiang et al proposed a method of forecasting differential global position system (DGPS) solutions between the current and previous moment by using the IMU velocity and heading as inputs, which proved that feed-forward neural networks with back propagation was capable of integrating measurements from an IMU and DGPS [13]. Song et al developed an algorithm which integrated IMU specific force and angular velocity increments as linear velocity and angular velocity increments to form the inputs, and a radial basis neural network was used to predict GPS displacement increments, significantly eliminating anomalies in the filtering state [14]. Abdel-Hamid et al constructed an inputoutput pattern where the inputs to the system were the IMU raw measurements, the attitude angles determined through mechanization and the GPS outage times, and the prospective outputs of the model were position errors during GPS signal outages, providing reliable position error estimates during the absence of the reference position measurements [3].…”