| One of the major challenges in current military commanding and decision support domain is the so-called"the Informationc Gap"between the"Force Level"information that commanders needed and the"Object Level"information provided by modern battlespace surveillance systems based on the first level data fusion technology. It leads to high cognitive pressure on military commanders. To bridge this gap and provide commanders with aggregated"Force Level"information, this paper systematically studies the problem of force aggregation.Based on related researches, firstly, corresponding definitions are clarified and a framework model surrounding around the"Group"is brought forward, which is an important element of battlespace situation. Then, some functions in this model are studied from technology perspective.The main contributions of this paper are as follows:1. Proposing a force aggregation framework model surrounding around"Group"Based on the SAW model provided by Endsley, this paper studies the connotation of SAW in military context combined with the military dicision making process at first, and clarify the definitions of"situation","situation awarenss"and"situation element"etc. Then, according to the information requirements of military cmmanders in military decision making, corresponding functions of force aggregation are organized around"Group", and the framework model of force aggregation is provided.2. Proposing the SFG (Single Frame Based Grouping) object grouping method.In this paper, object grouping problem is viewed as a multi-hypothesis checking problem at first. Then, the SFG method based on generic algorithm and cased-based reasoning is proposed to resolve this NP-hard problem, and its convergence and rationality is verified.3. Developing the MFG (Multi-Frame Based Grouping) object grouping method.Based on SFG object grouping method, futherly considering the factor of time, this paper develop the MFG object grouping method, and model the association between hypothesis in multi-frame data using one-zero programming.4. Devising a group classification method based on the idea of Evidential Theory.According to the uncertainty and ambiguity of multi-souce intelligence and the ambiguity of viewpoint in corresponding knowledge, this paper devise a group classification method base on the idea of Dempster-Shafer theory, in which the intelligence is represented and combined with the idea of Dempster-Shafer theory, and the corresponding knowledge is modeled as type templates at first, then, the type of group is recognized based on the matching between combined intelligence and the templates. 5. Designing the method of formation recognition based on Hough Transform.In this paper, the pattern features of a formation composed of lines is modeled as a set of feature points based on Hough Transform, and the corresponding pattern class is modeled as formation templates made up of a set of feature windows at first, then, the type of a formation is recognized based on the matching between its feature points and formation templates. |
