The Effects of Using Heading Measurement During Alignment of a Low-cost IMU/GPS System

M. Choi, D. Won, S. Sung, Y. Lee, J. Kim, J-P. Park, H-W. Park, Y.J. Lee

Abstract:	This paper presents a performance analysis of an alignment using GPS heading and magnetic sensor information outside coupled systems for low-cost IMU/GPS. An inertial navigation system is divided into gimbaled INS and strapdown INS. This system can calculate the position of vehicles and velocity using only the information from the gyro and accelerometer and without the aid of external information. SDINS has low power consumption compared with GINS and has the additional advantage of lower maintenance. SDINS has been applied to various fields using a low-cost MEMS IMU to the development of MEMS. However, in low-cost MEMS IMU, the initial error is large and the long-term stability is low. Therefore, we have developed a system that combines an external sensor system and provides stable operation results, such as the GPS. However, the GPS update cycle is long. However, combining the GPS and INS produces a system that complements the strengths and weaknesses of both. When combining the low-cost IMU and GPS, it is important to estimate the initial position of the IMU. The first alignment process is divided into coarse alignment and precise alignment. Next, the vehicle is aligned using a gyrocompass loop, and the angle (heading) of the yaw on the horizontal axis is estimated by using the measured value of the gyro. The vertical axis is used to estimate the initial pitch and roll of vehicle with an accelerometer. If the vehicle not moving, a filter, such as the Kalman filter, is used to estimate the bias of the sensor. A precision alignment method is used to estimate the initial attitude. If an alignment is required when a vehicle is moving, the in-flight alignment operation is performed. In this case, the performance of the alignment process can be improved, and additional measurements of the external sensor can be obtained. In this paper, additional external sensor readings of the measure heading according to information calculated from the velocity measurements of GPS and heading measurements obtained from the magnetic sensor, based on magnetic north. If the initial calibration process is done in a magnetic environment, the heading measurements obtained from the magnetic sensor have the advantage of certain accuracy. If the vehicle speed is zero, the heading obtained from the velocity of the GPS does not give an accurate heading measurement. However, if the speed is constant, a more accurate heading measurement can be calculated. The observability of the model and the measurement error model systems in the process of alignment are not optimal. To change the state variable of this system error model, the improved precision of the alignment variable is estimated, thus increasing the observability matrix. Some studies to improve the alignment process have increased the observability matrix by measuring the position of the external sensor. A disadvantage is that when an additional estimate of the states is needed, a complex system configuration of the GPS antenna must be installed. In this paper, without reducing the number of variables, the system is implemented by using a low-cost GPS sensor, a low-cost magnetic IMU, and a single antenna. Depending on the speed of the vehicle, an algorithm is applied select the heading measurements. If the vehicle is not moving, an alignment of the high-precision measurement model consisting of a matrix can be performed by using the measurements of the magnetic sensor. If the vehicle is moving with constant velocity, an in-flight alignment operation constituting the matrix of the model determined can be performed by using the heading measurements of the GPS. The filter consists of the EKF. Set the state variables of the 12th, the state variable, precise alignment of the state in the process of stopping. Set the state variable of sort 15 times in service. If this configuration of the heading measurement model is used to measure the value, it will increase the observability matrix by one increment. The rank is increased to improve the performance of the alignment. The variables affect the estimation performance. The performance of the system alignment of heading measurements of the external sensor was verified in an experiment using a car at Konkuk University at 10:00 am, November 02, 2012. The inertial sensors of the analog device were ADIS16364, and magnetic sensors were HMC1053 (Honeywell). The GPS receiver was a Ublox LEA-5T, and the raw data receiver was a Ublox LEA-5S. For comparison of the receivers, we used a Novatel Propak-V3 for the GPS receiver, and a Honeywell HG1700-AG11 for the inertial sensor. Using the raw data, the time differenced carrier phase (TDCP) and Doppler measurements were used to estimate the speed of the GPS. This paper presents a method for improving the performance of the yaw axis in the initial attitude estimation by implementing a low-cost IMU in the alignment process. An additional external sensor was used for the measurements of GPS and heading magnetic sensor. The data were used to analyze the error model view of the EKF. The results showed that alignment was selected using a magnetic sensor and GPS heading measurements, depending on the speed of the vehicle. The experimental results were used to analyze the convergence of the initial error by adding a measurement of the value of the alignment heading, which affected the axis of the yaw.
Published in:	Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013) September 16 - 20, 2013 Nashville Convention Center, Nashville, Tennessee Nashville, TN
Pages:	2175 - 2179
Cite this article:	Choi, M., Won, D., Sung, S., Lee, Y., Kim, J., Park, J-P., Park, H-W., Lee, Y.J., "The Effects of Using Heading Measurement During Alignment of a Low-cost IMU/GPS System," Proceedings of the 26th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2013), Nashville, TN, September 2013, pp. 2175-2179.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In