Research On Task Allocation&Attack Effectiveness Evaluation Technique For UAVs Cooperatively Combating

In recent years, the UCAV (Unmanned Combat Aerial Vehicle) has caused much concern aroundthe world because of its outstanding performance in reconnaissance, anti-terrorism and especiallyhigh-tech local wars. Multi-UAV collaborative combat techniques have been rapdily developed as aresult of this trend, and show a new combat mode and technology. In this dissertation, themulti-UCAV cooperative combat task allocation techniques and the target damage assessmenttechniques based on image analysis are studied. The main work and innovations are as follows:Firstly, according to the development requirements of the multi-UCAV collaborative combat andits characteristics, task allocation process is divided into two different phases which are pre-allocationstage under static environment and re-allocation stage under unexpected environment. Through theanalysis of the situations of different stages, the mathematical model and the entire UCAV systemstructure formation of the cooperative combat system are established.Secondly, the optimum allocation results are more important than the real-time requirementsunder the task pre-allocation. The task pre-allocation for multi-type UCVA is described as amulti-objective optimization problem based on the optimal characteristics of the algorithm.Meanwhile, an improved elitist strategy of fast Non-dominated Sorting Genetic Algorithm (NSGA-II)is developed for searching the optimum solution. It provides auxiliary support for the furtherdecision-making of the decision-makers.Following, at the re-allocation stage, the real-time ability of task allocation algorithm is moreimportant than the optimization result in order to obtain a satisfactory decision. As a result of this, thedistributed task allocation approach based on contract net and Multi-Agent-System (MAS) ispresented. The algorithm extends the traditional contract net, and designs auction mechanism to solvethe task re-allocation in unexpected environment. The simulation results show that the proposedalgorithm can quickly achieve the multi-UCAV task allocation in a short time.Finally, the damage assessment method is studied based on image information of the target ofcombat before and after. The damage assessment criteria are developed mainly by extracting andanalyzing the changes of geometric and texture features of the target iamge. According to thecorresponding assessment criteria, the decision-makers can determine the damage degree of the target.