Comparison of methods for association and fusion of multi sensor data for tracking applications

Girija, G and Raol, JR (2001) Comparison of methods for association and fusion of multi sensor data for tracking applications. In: AIAA Guidance, Navigation, and Control Conference and Exhibit, 6-9 Aug 2001, Montreal, Canada.

Full text not available from this repository.

Abstract

In this paper, the two commonly used algorithms for association namely the Nearest Neighbour (NN) and Probabilistic Data Association (PDA) are applied for measurement too track association, estimation/filtering and fusion. The algorithms are first validated on simulated data of a target moving with constant velocity and tracked by two sensors with different measurement noise characteristics. Each of the sensors is equipped with a Kalman Filter which is essentially used to update the track states after associating the incoming measurement with the associating the incoming measurement with the using hierarchal sensor architecture to generate a single fused track. A comparison of the two methods is made in terms of various statistical performance measures for the filters. For this case of simulated data of a single target, multisensor scenario the performance of the NN Filter (NNF) is superior to that of PDA Filter (PDAF) when the measurement noise levels are high. For low measurement noise levels, the performance of the two filters is similar. The algorithms are then applied to generate a fused trajectory from real data of a moving object.

Item Type: Conference or Workshop Item (Paper)
Uncontrolled Keywords: Data association methods;Multi sensor data;Numerical simulation results
Subjects: MATHEMATICAL AND COMPUTER SCIENCES > Mathematical and Computer Scienes(General)
AERONAUTICS > Aeronautics (General)
Division/Department: Flight Mechanics and Control Division, Flight Mechanics and Control Division
Depositing User: M/S ICAST NAL
Date Deposited: 07 Feb 2008
Last Modified: 24 May 2010 09:55
URI: http://nal-ir.nal.res.in/id/eprint/4579

Actions (login required)

View Item