Information-Driven Search for Multiple Moving Targets

To search for multiple moving targets in ocean surveillance by space-based sensors, an Information-driven approach is developed based on information theoretic metrics including Kullback-Leibler discrimination and entropy. Use probability distribution to represent of target positions. Calculate information gain from target probability distributions between motion prediction and hypothetical observations, and select the sensing action yielding maximum information gain. Monte Carlo method is used to approximate target states for motion prediction and hypothetic state enumeration to decrease memory and calculation consumption when grid number of search region and targets´ number is large. Finally the effectiveness of the proposed approach is qualified by simulations.