Study On UAF Formation Control In Obstacle-Laden Environment

Unmanned Aerial Vehicles have been extensively used in various areas such as intelligent surveillance, Surveying and Mapping, Aerial photography , Remote sensing , Disaster warning,networking communication in board. And by virtue of the advantages of the UAV formation, it has been one of the hotspots recently. The past few years have witnessed the thriving interest with UAV in Units. However, huge technical gaps still exist between our current level and that of abroad, and new concepts, new ideas, and new principle oriented at the UAV are still rare. On the other hand, the UAV formation technique is the key technology in UAV application, which can be used to extend its usage and improve its efficiency.The UAV formation is the significant trend in cooperative control of UAV, it has broad prospects for development, the single UAV will be deployed on many occasions when completing a more complex task, because carrying equipment for a single UAV is limited. The UAV formation can make the task completed more efficient., it can be dispersed units carry equipment, will be a complex task split into several simple tasks, assigned to the formation of different UAV. The UAV formation can carry different equipment and cooperation to complete a single task that UAV can not complete the task, because it has features and advantages as 2001, The UAV formation technique has been"key technology"in The USA army, and lots of manpower and material resources have been carry out the system development, the UAV formation is important in the future battlefield.The application performance of the UAV formation increased, it not only has single UAV and load getting better and better and more powerful, increasing reliability, while the overall need for collaboration formation flying capabilities. The traditional UAV formation take cooperative Master-Slave communication / control architecture. It is characterized by unity for target selection by the host, task allocation and coordination of planning, then implementation of the master-slave assignments. The framework is simple, but the reliability, autonomy and poor, if the host were destroyed, wingman will lose decision-making and central control agencies, the UAV formation will be paralyzed. This constraint will be more serious in the next long-range coordinated attack: the host of the damage will mean the complete failure of the overall mission, once fleet command structure from the ground. In addition, the main computation and amount of information of the host appears geometrically growth with aircraft group size increasing, it is difficult to achieve effective distribution of tasks and real-time scheduling attack, it is difficult for his future and more complex battlefield environment large-scale coordinated attack fleet.To overcome the inherent shortcomings of the slave control framework, the advanced ring architecture in collaborative control fields can been applied to the UAV formation coordinated attack. Different from the master-slave architecture, non-hierarchical distributed communications / control configuration are applied in the framework, The airplanes of the group of aircraft attach through flexible communication / sensing topology, it can realize real-time online distributed computing and communications, therefore, the requirements are reduced greatly in information processing, it is very easy to use simple,low-cost attack aircraft coordinated attack.Based on the background of the UAV formation, in this paper, the formation control scheme, and the formation and obstacle avoidance control for both the ideal and practical dynamic models are studied. We focus our attention on the follows three aspects:(1) Investigation on the UAV formation control is thoroughly conducted, and the decentralized scheme is proposed. The convenient architecture adopted is the"leader-follower"which is easy to implement but possesses several drawbacks, including poor reliability, scalability etc. Furthermore, as the scale of formation increases, great computation and communication overhead will insure be introduced, which may even cause the total failure of the whole system. The decentralized architecture will overcome this problem. According to the property of the formation control, in this paper, we divide the control into the high-level control and the low-level control, and we focus on the first one.(2) For ideal UAV dynamics, the control strategy for UAV formation with obstacle avoidance is designed and numerical simulations are provided to demonstrate the effectiveness of the proposed algorithm. Techniques including the artificial potential function and gyroscopic force are adopted in the control algorithm. Various of formation control strategies are compared and analyzed, and the rigid mathematical description of the problem is given. Graph theory is introduced and the communication and control is therefore combined. Through investigation, we find that the traditional APF method is easy to cause the local minimum points, and we introduce the gyroscopic force into our control algorithm. Therefore, more intelligent and practical trajectories are generated which is easy to implement. Distinguished from other literatures, our algorithm is not only capable of tracking constant trajectories but also the time-varying ones. Simulation results demonstrate the effectiveness.(3) Considering the practical dynamics of UAV as well as the model uncertainty and external disturbance, the UAV formation control algorithm based on MPC scheme is proposed. Numerical simulation is also provided to illustrate the effectiveness of this method. The detailed idea is as follows: by introducing the constraint function into the MPC architecture, we transform the problem into a optimization problem. The decentralized MPC controller design is studied in detail and under the condition that not too much performance is scarified, the decentralized control is implemented. In addition, the decentralized robust MPC is also designed and the robustness is guaranteed.Our future work involves the high-fidelity modeling of the UAV formation. currently, though controllers are developed, we oversimplified the model. In addition, the time-delay effects and the dropout properties are also required to be considered. And the evaluation of the overhead of the algorithm computation is also needed for practical engineering implementation.