A Grouping Method of Sea Battlefield Targets Based on The Improved Nearest Neighbor Algorithm

In the situation analysis and assessment aspect, the group identification of battlefield targets is very important. It is one of the key issues that commanders are concerned about in commanding operations, and it is also a necessary part of identifying the enemy's combat intentions. Considering inaccurate grouping phenomena of special patterns (non-“near-circle” patterns) that appeared on the battlefield, this paper adopts the improved nearest neighbor method to group sea battlefield targets. Firstly, the spatial features and attribute features of fusion targets are obtained, and then both the Euclidean distance (used to measure spatial similarity) and Mahalanobis distance (used to measure attribute similarity) between targets are calculated. Secondly, based on these two distance parameters, an improved nearest neighbor algorithm is used to finish the target grouping. Finally, update group parameters and maintain the addition, splitting, and missing changes of groups. The simulation results show that this improved algorithm can effectively accomplish the target grouping tasks of different patterns rising on the sea battlefield, and provide an effective basis for further battlefield situation understanding.