Bearing Only Track to Track Association for Airborne Surveillance System

This paper proposes robust bearing data-based association method for airborne sensor tracks. This method adopts existing algorithms namely, Probabilistic Neural Network (PNN) & Global Nearest Neighbour (GNN) for efficient association. In this method, modified average of instantaneous outcome from the PNN/GNN algorithm, confidence of previous update, and number of successive hits are considered for calculation of final confidence. This method is validated using sparse, crossing, and dense target related to combat scenarios in simulated environment. PNN/GNN algorithms performs poorly in dense and crossing target scenarios. In proposed method, we have added averaging scheme post PNN/GNN outcome to increase association accuracy.