of Surveying and Spatial Information Systems
University of New South Wales
Surveying Applications of Inertial Navigation Systems:
Experiments and Data Analysis
by Steve Hewitson
Supervised by Dr. Jinling Wang
Edited by J. M. Rüeger
Current inertial technology applications in the surveying industry include airborne and land-based mobile mapping systems (MMS), airborne gravity surveys, hydrographic and subsurface mapping, land surveys (such as cadastral and geodetic control), seismic surveys and other special applications. Aspects of data production, processing, including Kalman filtering, and analysis are discussed with reference to field experiments with a Boeing C-MIGITS II integrated GPS/INS system.
Experimental System Setup
The GPS/INS system was mounted on a stable platform on the roof of a vehicle, as shown in Figures 1 and 2. The sampling rates were 1 and 100 Hz for the GPS and IMU measurements, respectively. The GPS roving antenna received the measurements for updating the INS whilst the micro tracker antenna was used to synchronise the INS measurements to GPS time.
2: GPS/INS experimental platform layout
GPS/INS Hardware and Software
The Boeing C-MIGITS (Coarse/acquisition (C/A) code, Miniature Integrated GPS/INS Tactical System) II was released in December 1996. The C-MIGITS II consists of a Digital Quartz Inertial Measurement Unit (DQI IMU) and a 5-channel micro tracker low power commercial GPS receiver engine. The software used to process the data (AIMS) was originally developed by the Center for Mapping at The Ohio State University for 4000ssi GPS receivers and a strapdown Litton LN-100 INS. The University of New South Wales has developed a modified version of the AIMS processing software to suit the C-MIGITS II system.
The AIMS software was used to resolve the GPS ambiguities and to combine the double-differenced L1 carrier-phase measurements with the IMU measurements in a single Kalman filter. MATLAB was used to visualise the Kalman filter outputs and for binary/ASCII file format conversions. A Kalman filter was used to estimate 18 error states for the inertial navigation solution and IMU measurement errors, namely 3 position error states, 3 velocity error states, 3 attitude (pitch, roll and heading) error states, 3 accelerometer biases, 3 accelerometer scale factor errors and 3 gyro drifts.
Description of Experiments
Three experiments were conducted to test the system under low dynamic, high dynamic and urban conditions. The low dynamic test involved an oblique rectangular trajectory with velocities of about 6 m/s and a mission length of about 800 seconds. The high dynamic test featured vehicle velocities of up to 15 m/s and had a mission length of about 650 s. The experiment for urban conditions was conducted within an environment, where trees, buildings, power lines and other features obstruct satellite signals from being received by the roving GPS receiver. The length of the mission was about 1100 s. The typical velocity of the vehicle was between 10 and 15 m/s for this test. A 20 s blockage of the GPS signals was simulated, where the positioning solution is obtained from stand-alone INS.
Figure 3: Effect
on position of a 20 second GPS signal outage
The results show that the Kalman filter error estimates are of better quality, and stabilise much more quickly, for the high dynamic case than for the low dynamic test. The test equipment proved capable of providing a navigation solution in a normal urban environment, where features such as trees, power lines and buildings may obstruct GPS signals. A 20 s GPS outage, during which the position updates are obtained from the INS alone, caused an average error drift of 0.5 m and 0.4 m for horizontal and vertical position components, respectively.
For more information, please contact:
Dr. J. Wang
School of Surveying and Spatial Information Systems
University of New South Wales
UNSW SYDNEY NSW 2052