14 Scientific Technical Review, 2014,Vol.64,No.2,pp.14-20 UDK: 656.61.052:621.39 COSATI: 17-02, 17-07, 14-02 Adaptive Error Damping in the Vertical Channel of the Ins/Gps/Baro-Altimeter Integrated Navigation System Vlada Sokolović 1) Goran Dikić 1) Rade Stančić 2) In the inertial navigation system (INS), the altitude error diverges exponentially, especially in low-cost sensors. To suppress divergence of the altitude error, a Global Positioning System (GPS) receiver and a barometric altimeter (baro-altimeter) are utilized. This paper describes the error dumping of the vertical channel in the integrated navigation system INS/GPS and the baro-altimeter by using the 3 rd order vertical channel damping loop and the application of the adaptive error damp coefficients. The integration is done with an extended Kalman filter (EKF) with control signals. The characteristics of the proposed model for the integration of the INS, the GPS and the baro-altimeter were analyzed by a computer simulation as well as experimentally, using a vehicle. The results of the analysis show that the navigation solutions of the INS/GPS/Baro- altimeter navigation system could improve the accuracywith the adaptive error control coefficients in the EKF control signal. Key words: inertial navigation, navigation system, global positioning system, barometric altimeter, kalman filters, error correction, algorithms. 1) University of Defense, Military Academy, Generala Pavla Jurišića Šturma 33, 11000 Belgrade, SERBIA 2) Serbia and Montenegro Air Traffic Services SMATSA, 11000 Belgrade, SERBIA Introduction N inertial navigation system is an autonomous system for determining the position, velocity and attitude of an object using three linear accelerometers and three rate gyroscopes, [1]. In this paper, the "Strap-down" INS (SDINS) is used for analyzing and testing. The main characteristic of the low cost INS is its low accuracy. Weak points of these sensors are complex stochastic errors that are difficult to model, [2]. The vertical channel error diverges exponentially, so if the error is not dumped, the vertical data of the INS cannot be trusted during a long time period, [3]. The addition of baro-altimeters to integrated navigation systems is an already known idea, as they are typically inserted into INS systems to reduce error growth in the local vertical channel. The most widely used methods are a vertical channel damping loop (baro-inertial damping loop) and Kalman filter mechanization. So far, many researchers have focused on INSs constructed with low cost sensors in order to improve their accuracy and reduce total costs of navigation systems, [4]. Jaewon et al. proposed an error compensation model for the INS vertical channel, where the well-known vertical channel damping loop is constituted, and then with a GPS and the loop, a Kalman filter for error compensation is designed, [3]. Salychev et al. used GPS data to correct the IMU, [5]. Kim J.H. et al. presented the results of improving an INS/GPS navigation system with a baro-altimeter for an Unmanned Aerial Vehicle [6]. The results showed that the INS/GPS/Baro-altimeter navigation system could provide a more reliable and accurate navigation solution under high maneuvering environments by using a ten-state Kalman filter to blend the INS optimally with a GPS and a baro-altimeter, [6]. Satheesh Readdy et al. have shown that the baro-altimeter data and the INS data fuse with a four state Kalman filter to obtain an accurate estimate of the flight altitude, [7]. Sokolovic et al. used a 2 nd order vertical channel loop and an EKF with an adaptation of the vertical velocity component to improve the accuracy of the altitude, [8]. The aim of this work is to show that the application of the adaptive error (gain and damp) coefficients in the 3 rd order vertical channel loop of the integrated navigation system improve the accuracy of the navigation solution by using different non-linear functions for the adaptations of the control signals introduced in the EKF filter. Error model for the baro-altimeter The measurement model of the baro-altimeter includes the bias error of the 1 st order Markov process, the scale factor error of the random constant and white Gaussian noise [9], and is expressed in the next equation: , 1 , 0, B T B T T B h h h h B Sh B B w S δ ν τ = + = + + + =− + = & & (1) where, h B is the measurement of the barometer, h T is the true height, δh B is the error component of the barometer output, B is the bias error, S is the scale factor error, v B is the measurement noise, τ is the correlation time of the bias error, and w is driving white Gaussian noise for the bias error A