西北工业大学学报
西北工業大學學報
서북공업대학학보
JOURNAL OF NORTHWESTERN POLYTECHNICAL UNIVERSITY
2015年
1期
26-32
,共7页
邵壮%祝小平%周洲%王彦雄
邵壯%祝小平%週洲%王彥雄
소장%축소평%주주%왕언웅
无人机%编队飞行%队形反馈%虚拟结构%非线性模型预测控制
無人機%編隊飛行%隊形反饋%虛擬結構%非線性模型預測控製
무인궤%편대비행%대형반궤%허의결구%비선성모형예측공제
为提高编队大机动时的队形保持能力,采用虚拟结构编队方法,基于李雅普诺夫直接法设计独立的非线性队形保持控制器,并在此基础上采用非线性模型预测控制方法设计含队形反馈的编队轨迹跟踪器。通过在代价函数中引入队形误差代价来实现队形反馈控制策略,并采用动态参数实现编队队形保持和沿参考轨迹飞行之间的自适应切换,各无人机通过滚动求解有限时域优化问题得到虚拟结构的控制指令。仿真结果表明,相较于无队形反馈的情况,所设计的含队形反馈轨迹跟踪器能够显著地降低编队大机动时的队形保持误差。
為提高編隊大機動時的隊形保持能力,採用虛擬結構編隊方法,基于李雅普諾伕直接法設計獨立的非線性隊形保持控製器,併在此基礎上採用非線性模型預測控製方法設計含隊形反饋的編隊軌跡跟蹤器。通過在代價函數中引入隊形誤差代價來實現隊形反饋控製策略,併採用動態參數實現編隊隊形保持和沿參攷軌跡飛行之間的自適應切換,各無人機通過滾動求解有限時域優化問題得到虛擬結構的控製指令。倣真結果錶明,相較于無隊形反饋的情況,所設計的含隊形反饋軌跡跟蹤器能夠顯著地降低編隊大機動時的隊形保持誤差。
위제고편대대궤동시적대형보지능력,채용허의결구편대방법,기우리아보낙부직접법설계독립적비선성대형보지공제기,병재차기출상채용비선성모형예측공제방법설계함대형반궤적편대궤적근종기。통과재대개함수중인입대형오차대개래실현대형반궤공제책략,병채용동태삼수실현편대대형보지화연삼고궤적비행지간적자괄응절환,각무인궤통과곤동구해유한시역우화문제득도허의결구적공제지령。방진결과표명,상교우무대형반궤적정황,소설계적함대형반궤궤적근종기능구현저지강저편대대궤동시적대형보지오차。
When executing a sharp maneuver, if the formation maneuver path is not well designed, a formation of UAVs constrained by their performance capabilities, may not be able to keep in satisfactory formation as desired. We,applying Lyapunov stability theory and using virtual structure approach , propose designing individually a non-linear controller for satisfactory formation keeping .Then,by applying nonlinear model predictive control approach , a trajectory tracker with formation feedback is designed to improve the capacity of formation keeping during a sharp maneuver.Formation feedback is achieved by adding a formation-error cost to the total cost function,and a dynamic parameter is proposed to obtain a trade-off between formation keeping and flying along prescribed trajectory adap -tively.Each UAV gets the reference commands of the virtual structure by solving a Finite Horizon Optimal Control Problem at each time step.Simulation results and their analysis suggest that, compared with situation without forma-tion feedback, the designed trajectory tracker with formation feedback can reduce formation-keeping error evidently.