ANN-assisted robust GPS/INS information fusion to bridge GPS outage

Mehdi Aslinezhad, Alireza Malekijavan, Pouya Abbasi
2020 EURASIP Journal on Wireless Communications and Networking  
Inertial navigation is an edge computing-based method for determining the position and orientation of a moving vehicle that operates according to Newton's laws of motion on which all the computations are performed at the edge level without need to other far resources. One of the most crucial struggles in Global Positioning System (GPS) and Inertial Navigation System (INS) fusion algorithms is that the accuracy of the algorithm is reduced during GPS interruptions. In this paper, a lowcost method
more » ... for GPS/INS fusion and error compensation of the GPS/INS fusion algorithm during GPS interruption is proposed. To further enhance the reliability and performance of the GPS/INS fusion algorithm, a Robust Kalman Filter (RKF) is used to compensate the influence of gross error from INS observations. When GPS data is interrupted, Kalman filter observations will not be updated, and the accuracy of the position will continuously decrease over time. To bridge GPS data interruption, an artificial neural network-based fusion method is proposed to provide missing position information. A well-trained neural network is used to predict and compensate the interrupted position signal error. Finally, to evaluate the effectiveness of the proposed method, an outdoor test using a custom-designed hardware, GPS, and INS sensors is employed. The results indicate that the accuracy of the positioning has improved by 67% in each axis during an interruption. The proposed algorithm can enhance the accuracy of the GPS/INS integrated system in the required navigation performance. standard navigation systems in transportation applications. Inertial Navigation System (INS) consists of 3-axis sensors measuring linear acceleration and angular velocity in order to calculate position, velocity, and Euler angles. With enhancements in processors, integration of two or more low-cost sensors with a more complex but accurate fusion algorithm can have a significant influence on the extension of flight, shipping, and car industries [2, 3] . For GPS/INS integrated system, the navigation solution is traditionally achieved by using Kalman Filter (KF). However, the major inadequacy related to KF is that the system states are usually nonlinear. Extended Kalman Filter (EKF) is the nonlinear version of the KF, which linearizes about an estimate of the current mean and covariance. Since there are two integrators in calculating position from acceleration data, a second-order EKF is the most suitable approach in data fusion algorithms [4, 5] . In order to improve the accuracy of the position and compensate the INS error in the long-term absence of GPS data, the neural network is an excellent approach to reduce the cumulative error of INS [6]. In this paper, the ultimate objective is to train an artificial neural network from GPS data to learn the latest available position signals and then use this trained network to predict the time series in the absence of the GPS signal. Considering extensive capability of neural networks to solve nonlinear problems and overcome the drawbacks of KF and different types of Artificial neural networks (ANN) such as Multilayer Perceptron (MLP), Radial Basis Function (RBF), and Adaptive Neuro-Fuzzy Inference System (ANFIS) approaches are used in data fusion applications [7, 8] . Also, a variety of training methods are used in this field, but the most common methods are based on error Back-Propagation (BP) methods [9, 10] . Recently, there has been a significant focus on GPS and INS sensor data fusion algorithms, especially in absolute position estimation during a failure in communication with sensors or an outage in GPS. The primary focus in most of the articles is to maintain position accuracy improvement. For example, in [11] , Design of an INS/GPS algorithm using parallel Kalman filters is investigated. In [12] , an Adaptive IIR/FIR fusion filter approach is implemented. In [6, 13], a GPS/INS data fusion is applied using Unscented Kalman Filter (UKF) and RKF. In [14], the Fusion algorithm using UKF and BP neural networks is investigated. Also, in [15, 16] , hybrid integration methods using different types of neural networks are investigated. Kalman Filter is the most remarkable real-time optimal estimator, which has found a vast field of application [17] . KF has become an indispensable data fusion approach for GPS/INS integrated navigation. Compared to other alternative methods of GPS/INS fusion like Particle Filter [18] or IIR/FIR Filers [19] , Kalman Filter proposes a better adaptivity and flexibility of implementation in low-cost systems along with strong optimality and accuracy of estimation. However, the standard KF algorithm can hardly deal with nonlinearity in model and robust issues such as uncertainty. In this research, EKF has been utilized to solve the issue of dealing with nonlinearity. Recently, researchers have proposed and suggested robust filter algorithm [20] [21] [22] which stands for high stability and flexibility and better capability of controlling the negative effect of both the dynamic model uncertainty and the long-term measurements errors. The performance of integrated positioning will be degraded due to the gross error from observations in challenging
doi:10.1186/s13638-020-01747-9 fatcat:w5hyz747efgbliyncr76krgohe