Extraction And Application Of Decision Rules Based On Local Perception

For the conventional decision system, it often makes a decision after collecting all the relevant target information first. However, the decision system doesn’t give the perceiving guiding in the process of collecting information and making a classification decision, in another way, it can’t complete the classification and recognition quickly by full of using the known information. Therefore, according to the knowledge of rough set theory, this article research the algorithm which puts known information and perceiving guiding into the decision rule, and this article also achieves the target recognition and multi UAV(Unmanned Aerial Vehicle, UAV) cooperative reconnaissance based on the decision rule of local perceiving, the main work is as follows:Firstly, proposing the algorithm of decision rule based on local perceiving(DR_LP), and the algorithm of perceiving and searching decision rule(PSDR) based on DR_LP, matching the algorithm of DR_LP rules.1. The condition attributes’ position of conventional decision rule is not clear, and it doesn’t have the role of the direction guidance. Based on the situation, the algorithm of DR_LP is proposed. In the condition of knowing some information, by introducing the attribute importance of other conditions attribute information referring the known information relative to decision attribute, and then traversing all known information, it can achieve results which makes the known information and the perceiving and guiding into the decision rules.2. The algorithm of PSDR not only introduces mutual information achieving the known information by selecting the optimum information from multi-information, but also it achieves quickly searching corresponding decision rules by the way of mapping tables when some of the information is known. The combination of PSDR and DR_LP makes the decision system achieving that “Perceiving partial information Searching decision rules guiding and perceiving new information Achieving classification”, and it also improves the efficiency of searching information and making classification.Secondly, constructing the recognition model of decision rules based on local perceiving. In the conventional pattern recognition method based on the knowledge of rough, putting forward one pattern recognition model which contains realizing the continuous data’s discretization through FCM(Fuzzy C-means: FCM), attribute ®reduction based on Pawlk attribute importance, decision rules based on DR_LP and the classification and recognition based on PSDR. At last, by the use of experimental it verifies that the model has a high rate of complete information and certain recognition rate of incomplete information.Thirdly, the multi UAV cooperative reconnaissance based on DR_LP. In the conventional multi UAV cooperative reconnaissance based on snake searching, DR_LP is brought in, and each UAV reconnoiters in different area firstly, and then planning the UAV track through the DR_LP rules. The experimental results show that the multi UAV cooperative reconnaissance based on snake search and DR_LP shortens the distance of UAV reconnaissance, reducing the cooperative reconnaissance time, and finally the efficiency of multi UAV cooperative reconnaissance is greatly improved.