A novel parallel constrained extended Kalman filter for improving navigation algorithm – case study: gas pipeline

Abstract

In many real navigation problems, moving objects may have some state or measurement constraints along their way. Using these constraints in conventional Extended Kalman Filter (EKF) equations results large matrices which are computationally time-consuming. In this paper, a new Constrained Navigation Filter (CNF) is proposed as a parallel to reduce the computational burden of the conventional EKF algorithm while increasing the positioning accuracy. So a methodology has been developed for Strap-down Inertial Navigation System (SINS) based on MEMS IMU applied on Pipeline Inspection Gauge (PIG) to sense data at constant sampling rate of 108 km of the pipeline. The results verified that using such a hybrid approach has improved positional accuracy 8.97% in comparison with that of the latest methods like EKF/ Pipe Line Junctions (PLJ). Also, the proposed method is 2.277 times better than EKF/PLJ in the algorithm runtime.

Keywords:

• EKF
• MEMS IMU
• Odometer
• Navigation constraint
• PIG
• PLJ
• SINS

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

I. Hatefi Afshar

I. Hatefi Afshar is a Ph.D. student at the University of Tehran, Iran, GIS department, school of surveying and geospatial engineering. He received a bachelor's degree in surveying engineering in Tehran University and a master's degree in Geospatial Information Systems (GIS) from Tehran University in Tehran, Iran. His current field placement is with the National Iranian Gas Company (NIGC). He is interested in GIS, navigation solutions, data fusion, and geospatial data analysis.

M. R. Delavar

M. R. Delavar has obtained a B.Sc. in Civil Eng.-Surveying from KNT University, Iran in 1988, a MSc. in Civil Eng.-Photogrammetry and Remote Sensing from University of Roorkee (currently IIT Roorkee), India in 1992 and a PhD in Geomatic Eng.-GIS from the University of New South Wales, Australia in 1997. His research interests are in spatial data quality and uncertainty modeling, spatio-temporal GIS, spatial data fusion, spatio-temporal data mining, remote sensing and GIS integration.

B. Moshiri

B. Moshiri received his B.Sc. degree in mechanical engineering from Iran University of Science and Technology (IUST) in 1984 and M.Sc. and Ph.D. degrees in control systems engineering from the University of Manchester, Institute of Science and Technology (UMIST), U.K. in 1987 and 1991 respectively. He has been as adjunct professor of the Department of ECE at university of Waterloo and member of PAMI group since May 2014. His fields of research include advanced industrial control, data fusion theory and feasibility studies on applications and implementations of sensor/data, and Spatio-Temporal GIS.