摘要
随着无人机技术的逐渐成熟,制造成本不断降低,无人机在多个领域得到广泛应用,如农业、工业、安全、军事等.而无人机在执行任务时,往往需要与地面基站保持通信的连通性.但在距离较远或有遮挡情况下,信道质量将严重下降,这对通信连通性的保持提出挑战.针对该问题,本文考虑部署多架无人机作为通信中继,以串联方式构建任务无人机与地面基站稳定的通信链路.本文以长期信道容量作为优化指标,基于动态规划理论提出了两种中继无人机的规划方法:CMMP-AT和CMMP-OBO. CMMP-AT方法强调在任务执行初期就部署全部中继无人机,该方法部署较为简便,但计算复杂度较高. CMMP-OBO方法提出按照任务需求逐架部署中继无人机,该方法较为灵活,且计算复杂度低,扩展性好,可节约中继无人机的运动能耗.实验结果表明, CMMP-AT方法针对一个两中继无人机场景,需要40.03 h规划结果,而CMMP-OBO方法只需57.66 s即可规划出结果,并且可节约3.87%的运动能耗.此外,为了精确控制中继无人机遵循规划出的轨迹行进,本文基于模型预测控制方法实现对规划轨迹的追踪,并在V-REP环境中实现了多中继无人机场景的仿真.仿真结果表明,相比PID控制方法,模型预测控制方法能够更精确地追踪规划轨迹.在两种仿真场景下,模型预测控制方法的追踪误差仅为PID控制方法的43.97%和41.42%.
With the continuous development of unmanned aerial vehicle(UAV) technology and its falling cost, UAV has been widely used in important fields, such as agriculture, industry, security, and military. In operation, UAVs often need to maintain communication connectivity with ground stations. However, in the case of long distance or shadowing, the quality of the wireless channels will be seriously degraded, which will affect the communication connectivity. To solve this problem, this paper considers deploying multiple UAVs as communication relays. In this paper, we select the long-term channel capacity as the optimization objective and propose two planning methods for UAV relays based on dynamic programming(DP): CMMP-AT and CMMP-OBO. The CMMP-AT method emphasizes the deployment of all relay UAVs at the beginning of the task. This method is simpler to implement, but the computational complexity is intolerable. The CMMP-OBO method proposes to deploy relay UAVs one by one according to the task requirements. Although the CMMP-OBO method may lose performance, it has a lower computational complexity and better scalability and can save more motion energy. The experimental results reveal that the CMMP-AT method requires 40.03 h to obtain the planning results for a two-relay scenario, whereas the CMMP-OBO method only needs 57.66 s and can save 3.87% motion energy. In addition, to precisely control the relay UAVs to follow the planned trajectory, this paper implements trajectory tracking,based on model predictive control(MPC), and realizes the simulation of the multi-relay UAV scenario in the VREP environment. The simulation results reveal that compared with the PID control method, the MPC method can track the planned trajectory more accurately. In two simulation scenarios, the tracking error of MPC is only 43.97% and 41.42% of the PID control method.
作者
武云龙
张博
任小广
王彦臻
易晓东
WU YunLong;ZHANG Bo;REN XiaoGuang;WANG YanZhen;YI XiaoDong(National Innovation Institute of Defense Technology(NIIDT),People’s Liberation Army(PLA)Academy of Military Science,Beijing 100071,China;Tianjin Artificial Intelligence Innovation Center(TAIIC),Tianjin 300457,China)
出处
《中国科学：技术科学》
EI
CSCD
北大核心
2020年第5期538-550,共13页
Scientia Sinica(Technologica)
基金
国家重点研发计划(编号:2017YFB1001900)
国家自然科学基金(批准号:91648204,61906212,61601486,61802426)资助项目。
关键词
通信连通性
动态规划
多无人机系统
通信中继
模型预测控制
communication connectivity
dynamic programming
multi-UAV system
communication relay
model predictive control
