Seamless Navigation via Dempster Shafer Theory Augmented by Support Vector Machines

Deepak Bhatt, Priyanka Aggarwal, Vijay Devabhaktuni, Prabir Bhattacharya

Abstract:	In any military/civilian application—be it on ground, air or underwater—a sense of navigation (consisting of position, velocity and attitude parameters) is critically important. Global Positioning System (GPS) have been the prominent technology to fulfill this demanding requirement for reliable navigation over extended periods of time, covering any part of the world during day or night. However, in a number of scenarios, standalone GPS measurements may be completely unreliable (like in canyons or forests) due to the signal blockages or multipath effects etc (in urban areas). Under these circumstances, alternative information sources, like Inertial Navigation System (INS), need to be employed to bridge the gap of times when GPS signal reception is unavailable. In this paper we implement a novel hybrid methodology based on Dempster Shafer theory (DS) augmented by Support Vector Machine (SVM) to integrate INS and GPS measurements; thereby enhancing the accuracy of the low-cost Micro-Electro-Mechanical-Systems (MEMS) based navigation system. Traditionally, Kalman Filtering (KF) based approaches have been developed to fuse the INS and GPS data. KF is an optimal filter for linear systems with Gaussian noises but is not applicable to non-linear systems. For non-linear models, EKF (i.e., linearized KF) can be implemented but may cause filter divergence under high dynamic conditions. Alternatively, Particle Filtering (PF) and Artificial Neural Network (ANN) based approaches have been proposed. The major challenge of basic PF is the computational load, while ANN suffers from local minimization and overfitting problems. Hence, in order to effectively fuse GPS and INS data for land vehicle navigation application, we propose an efficient hybrid methodology based on Dempster Shafer theory augmented by Support Vector Machine (DS-SVM). Our field test results clearly indicate that the proposed DS-SVM algorithm effectively compensated and reduced positional inaccuracies over the regular ANN model during GPS availability and outage conditions for low-cost inertial sensors.
Published in:	Proceedings of the 25th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2012) September 17 - 21, 2012 Nashville Convention Center, Nashville, Tennessee Nashville, TN
Pages:	98 - 104
Cite this article:	Bhatt, Deepak, Aggarwal, Priyanka, Devabhaktuni, Vijay, Bhattacharya, Prabir, "Seamless Navigation via Dempster Shafer Theory Augmented by Support Vector Machines," Proceedings of the 25th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS 2012), Nashville, TN, September 2012, pp. 98-104.
Full Paper:	ION Members/Non-Members: 1 Download Credit Sign In