Abstract: Aiming at the problems existing in fuzzy reasoning track association, such as the filter divergence, the membership performance degradation, the difficulty in determining association threshold value, and the association effect deterioration under complex circumstances, a track association algorithm based on modified fuzzy theory and D-S Evidence Decision is proposed. Based on the attenuated memory extended Kalman filter (AMEKF), a logarithmic likelihood function adaptive correction method is proposed to correct the disorientation of membership function,and a D-S evidence decision-making tracking association algorithm with conflict resolution and consistent focusing is proposed. The simulation results show that comparing with the original algorithm, the new algorithm improves the membership performance and lowers the difficulty in determining the correlation threshold value. This new association algorithm improves the accuracy of track average Association decision by 5.3%.

Key words: fuzzy synthetic decision, logarithm likelihood, membership function, correlation threshold, collision resolution, D-S evidence decision