The Research Of Data Fusion Perception Rule Mining In UAVs Cooperative Network

In order to make UAVs lasting flight more time with the limited load capacity, the number of sensors equipped in one UAV is limited. However, in order to obtain more abundant information in the scene, so a plan that some UAVs with different kinds of sensors detect the targets cooperatively is applied. Since sensor resources are limited, the UAVs select which sensors to perform the task, how to make different types of sensor work together effectively and realize the advantage of multiple sensors, and how to avoid the waste of sensor resources are the key problems to be resolved.In this paper, a target recognition decision table is built which can fuse the data detected by different sensors based on Rough Set theory. This target recognition decision table is formed by find the target identified rules from the data set. In order to improve the efficiency of multiple source sensors, this paper puts forward the Sensors Management Framework based on the Perceptual Decision-Making Rule guide. Through the multiple source sensors data fusion perception rule mining, the detection mission of the UAVs arranged under this rule can realize the best detection efficiency with the minimum sensor resources in the UAVs cooperative detection network.Firstly, this paper introduce the fuzzy set and rough set theory in the detail. After analyzing some reduction algorithm of the fuzzy-rough set, this paper design the number of clustering adaptive fuzzy c-means clustering algorithm of target feature information table blur clustering and improve the fuzzy-rough set Quick Reduction algorithm to get knowledge from the target data set and then a target recognition decision table is built. Then this paper study the significance of attributes perception rule mining algorithm and put forward a kind of Perceptual Decision-Making Rule algorithm based on dependence of local attributes. At last, detection missions are arranged to the UAVs by combining the Sensors Management Framework based on the Perceptual Decision-Making Rule guide and the detection efficiency of different sensors with part of the target information.