A Study Of Task Assignment For Unmanned Aerial Vehicles

Applications of unmanned aerial vehicles range from military operations, emergency response to disaster relief. Its convenience and flexibility bring significant development to our national defense and social life. Therefore, scholars from different countries regard these UAV-related researches as hot spots which need to be concerned and explored.Scholars have studied really further in the area of unmanned aerial vehicles, and they have found a lot of problems need to be solved and to be improved. This paper, a study related to task assignment technology of the unmanned aerial vehicles, proposes new solutions to those problems.In the problem of UAVs’ task assignment, many researchers only concern about one factor in this problem and make their decisions that can only act well in one circumstance but may act badly in other circumstances. This paper shows us the UAVs’ task assignment technology based on evaluation mechanism. This theory not only concerns about the factor of distance but also concerns about the factor of safety and so on. Before making its decision, each UAV will evaluate both the profits it can get and the price it has to pay. Thanks to that theory we can finally get a balanced optimal solution in which the UAV can get more profits with less cost. In the problem of UAVs’ joint combat, we can easily find that the combat units vary from each other and the battlefield is so complex and always changing at the same time. In order to win the combat, this paper proposed a solution to make decisions for each UAV in the battlefield. This solution uses a kind of table, coorperation table as called in this paper, which is full of UAVs’ information to make decisions for the UAVs and helps them choose which UAV can be cooperated with in the joint combat. Most importantly, we take all kinds of constraints into consideration and assign the joint UAVs their target that need to kill first, in which way we can make sure that the enemy UAVs can be killed as quickly as we can. Faced with the mission of solving the problem of UAVs’ reconnaissance task, standard ant colony optimization algorithm is easy to fall into local optimum, and the solution’s efficiency is not good enough, so this paper uses a new method which is improved on the standard ant colony optimization algorithm but with a changing volatility to solve the problems above. When the reconnaissance task is changed, the original method can’t handle this change, so this paper uses the heuristic Q-learning algorithm to deal with this problem and re-plans the flight path of UAVs to finish their tasks. Finally, the simulation results we have got from our simulation platforms prove that the theory we proposed above can solve problems like one-side decision, inefficient solution, local optimization and so on in the problem of UAVs’ task assignment. We have achieved good experimental results to support our theories in this paper.