|

Using Artificial Neural Networks to Compensate for the Error in an Integrated Navigation System

Authors: Al Bitar N., Gavrilov A.I. Published: 05.06.2020
Published in issue: #2(131)/2020  
DOI: 10.18698/0236-3933-2020-2-4-26

 
Category: Informatics, Computer Engineering and Control | Chapter: System Analysis, Control, and Information Processing  
Keywords: inertial navigation systems, global navigation satellite systems, neural networks, unscented Kalman filter

The paper presents a new method for improving the accuracy of an integrated navigation system in terms of coordinate and velocity when there is no signal received from the global navigation satellite system. We used artificial neural networks to simulate the error occurring in an integrated navigation system in the absence of the satellite navigation system signal. We propose a method for selecting the inputs for the artificial neural networks based on the mutual information (MI) criterion and lag-space estimation. The artificial neural network employed is a non-linear autoregressive neural network with external inputs. We estimated the efficiency of using our method to solve the problem of compensating for the error in an integrated navigation system in the absence of the satellite navigation system signal

References

[1] Grewal M.S., Weill L.R., Andrews A.P. Global positioning systems, inertial navigation, and integration. New York, Wiley, 2007.

[2] Hofmann-Wellenhof B., Lichtenegger H., Wasle E. GNSS --- global navigation satellite systems. Springer, 2007.

[3] Crassidis J.L., Junkins J.L. Optimal estimation of dynamic systems. New York, CRC press, 2011.

[4] Al Bitar N., Gavrilov A.I. SINS/SNS integration based on loosely coupled scheme using extended Kalman filter. Inzhenernyy zhurnal: nauka i innovatsii [Engineering Journal: Science and Innovation], 2019, no. 4 (in Russ.). DOI: http://dx.doi.org/10.18698/2308-6033-2019-4-1870

[5] Al Bitar N., Gavrilov A.I. Comparative analysis of fusion algorithms in a loosely-coupled integrated navigation system on the basis of real data processing. Giroskopiya i navigatsiya, 2019, vol. 27, no. 3, pp. 31--52 (in Russ.). DOI: http://dx.doi.org/10.17285/0869-7035.0001

[6] Wan E.A., Van Der Merwe R. The unscented Kalman filter for nonlinear estimation. Proc. IEEE AS-SPCC, 2000, pp. 153--158. DOI: https://doi.org/10.1109/ASSPCC.2000.882463

[7] Potapov A.S. Tekhnologii iskusstvennogo intellekta [Artificial intelligence technologies]. St. Petersburg, SPbGU ITMO Publ., 2010.

[8] Noureldin A., Osman A., El-Sheimy N. A neuro-wavelet method for multi-sensor system integration for vehicular navigation. Meas. Sc. Technol., 2003, vol. 15, no. 2, pp. 404--412. DOI: https://doi.org/10.1088/0957-0233/15/2/013

[9] Rojas R. Neural networks. A systematic introduction. Springer, 2013.

[10] Taha M.R., Noureldin A., El-Sheimy N. Improving INS/GPS positioning accuracy during GPS outages using fuzzy logic. Proc. 16th ION GPS/GNSS. Portland, 2003, pp. 499--508.

[11] Jang J.S. ANFIS: adaptive-network-based fuzzy inference system. IEEE Trans. Syst., Man, Cybern., 1993, vol. 23, no. 3, pp. 665--685. DOI: https://doi.org/10.1109/21.256541

[12] Noureldin A., Karamat T.B., Eberts M.D., et al. Performance enhancement of MEMS-based INS/GPS integration for low-cost navigation applications. IEEE Trans. Veh. Technol., 2009, vol. 58, no. 3, pp. 1077--1096. DOI: https://doi.org/10.1109/TVT.2008.926076

[13] Wang X., Chen J.X., Ni W. A hybrid prediction method and its application in the distributed low-cost INS/GPS integrated navigation system. 18th Int. Conf. Information Fusion. Washington, DC, 2015, pp. 1205--1212.

[14] Chen X., Shen C., Zhang W.B., et al. Novel hybrid of strong tracking Kalman filter and wavelet neural network for GPS/INS during GPS outages. Measurement, 2013, vol. 46, no. 10, pp. 3847--3854. DOI: https://doi.org/10.1016/j.measurement.2013.07.016

[15] Zhang J., Walter J.J., Miao Y., et al. Wavelet neural networks for function learning. IEEE Trans. Signal Process., 1995, vol. 43, no. 6, pp. 1485--1497. DOI: https://doi.org/10.1109/78.388860

[16] Brown G. A new perspective for information theoretic feature selection. Proc. 12th AISTATS. Vol. 5. Clearwater Beach, Florida, 2009, pp. 49--56.

[17] He X., Asada H. A new method for identifying orders of input-output models for nonlinear dynamic systems. American Control Conf., 1993, pp. 2520--2523. DOI: https://doi.org/10.23919/ACC.1993.4793346

[18] Xie H., Tang H., Liao Y.H. Time series prediction based on NARX neural networks: an advanced approach. Int. Conf. Machine Learning and Cybernetics, 2009, vol. 3, pp. 1275--1279. DOI: https://doi.org/10.1109/ICMLC.2009.5212326

[19] Jekeli C. Inertial navigation systems with geodetic applications. Walter de Gruyter, 2012.

[20] Emel’yantsev G.I., Stepanov A.P. Integrirovannye inertsial’no-sputnikovye sistemy orientatsii i navigatsii [Integrated inertial navigation satellite systems]. St. Petersburg, Kontsern "TsNII "Elektropribor" Publ., 2016.

[21] Matveev V.V. Inertsial’nye navigatsionnye sistemy [Inertial navigation systems]. Tula, TulGU Publ., 2012.

[22] More J.J. The Levenberg --- Marquardt algorithm: implementation and theory. In: Numerical Analysis. Springer, 1978, pp. 105--116.

[23] Karsoliya S. Approximating number of hidden layer neurons in multiple hidden layer BPNN architecture. IJETT, 2012, vol. 3, no. 6, pp. 714--717.

[24] Stathakis D. How many hidden layers and nodes? Int. J. Remote Sens., 2009, vol. 30, no. 8, pp. 2133--2147. DOI: https://doi.org/10.1080/01431160802549278