Research On UCAV Tactical Operation Sequence Planning Based On Graphplan

Pre-war planning for UCAV(Unmanned Combat Aerial Vehicles) executing ground attack tasks includes the behavior sequence planning, and under it the route centered planning. The former one is a typical intelligent planning problem. Using intelligent planning method to solve the UCAV ground attack tactical operation planning problem is a feasible idea. This paper attempts to obtain the tactical operation sequence by means of Graghplan method. The main works are as follows:(1) Presenting the UCAV ground attack tactical operation planning model based on the planning conceptual model. In the typical combat scenario for UCAV, according to the characteristics of the problem, a relaxation is made for the system static assumption, i.e. in different circumstances the effect of actions may be different, and therefore conditional planning method is adopted. In the light of three elements of conceptual model of planning problem the model is established combining with the analysis to the specific example of aircraft ground attack. Based on the concept of tactical operation, the model is simple, clear and easy to expand, successfully reflecting the course of tactical operation.(2) Application research of Graphplan method to the UCAV ground attack tactical operation planning problem has been studied. Graphplan algorithm is adopted to solve the problem. During the solution analysis, we find that a large number of attack operators have caused frequent computational failure and time consuming. Besides, there is unnecessary backtracking within the solution searching process. Grounded on this, the algorithm improvement starts with operator expression and solution searching in the way of combining the conditional effects expression with minimum remaining first strategy. On one hand, conditional statements are added into action effects, so that the action whose effect differs with condition could be expressed by only one operator, which greatly reduces the number of action instances; On the other hand, during the solution searching process the target proposition which can be achieved by least action is first chosen. For example, if a certain proposition can only be achieved by a unique action, it is the one that is first selected to solve. Experimental results confirm the effectiveness of the algorithm improvement. This improved Graphplan method can successfully output the UCAV ground attack tactical operation planning sequence to achieve the preflight plan.(3) A battle scenario is set to do the simulation analysis. In accordance with the problem description and problem model a typical combat scenario of UCAV ground attack tactical task is designed and a UCAV-Graphplan procedure is produced to realize the improved algorithm. Finally the UCAV ground attack tactical operation planning problem is solved and the operation sequence is gained. Simulation results show that the model and the improved algorithm are both effective and reasonable. Therefore, applying the Graphplan method to solve the problem of UCAV ground attack planning proves out to be a primarily feasible idea.