A task-driven multi-source information fusion method

Abstract

The invention provides a novel task-driven multi-source information fusion method. Aiming at the characteristics of multi-platform and multi-sensor networking, a multi-thread information fusion mechanism is proposed, which can be adjusted automatically according to tasks. Firstly, based on multi-platform task planning, the transformation mapping from different application tasks to situation information requirements is realized by multi-objective decision-making method. Then, the real-time detection data quality of the sensor is evaluated based on the multi-source comprehensive verification comparison. The quality assessment results are generated based on the computing and communication resource allocation of sensor network and the situation, and the information fusion capability under thecurrent state is predicted. Finally, the multi-task dynamic scheduling method is used to generate the optimized information fusion threads under multi-constraint conditions, and the automatic matchingand adjustment of the corresponding fusion threads can be realized under the dynamic and changeable task conditions. This method can provide support for the generation of reasonable and applicable situation information for multiple application platforms under the background of diversified tasks.

Description

technical field [0001] The invention belongs to sensor network information fusion technology, relates to the multi-target tracking problem in multi-platform multi-sensor network data fusion, and provides a new task-driven multi-source information fusion method. Background technique [0002] At present, in many military and civilian application fields such as battlefield situation monitoring, traffic control, border patrol, disaster emergency rescue, agricultural situation detection, and air search and tracking of specific ground targets, air, space, ground, sea-based and other platforms (such as satellite , early warning aircraft, unmanned aerial vehicles, shore-based radars, maritime ships, etc.) Situation monitoring of key areas of concern has become an important means of obtaining environmental and target information. However, in a complex application environment, especially the battlefield, due to various factors such as diverse tasks and limited time, limited platform r...

AI Technical Summary

Problems solved by technology

[0004] However, the traditional platform-centric information fusion model is mostly a local application link for a certain type of users, usually with one or several fixed task styles as the application point, and adopts a "chimney" hierarchical tree structure as the main body. The network performs information interaction and processing. The internal coupling degree of the system is strong, and the application mode is relatively fixed. It has the characteristics of multiple levels of information interaction, weak function permeability, poor dynamic adaptability, and low information fusion efficiency. It is difficult to support the multi-platform refinement situation. Information Assurance

Images