摘要
随着人类航天活动的增加,空间任务正朝着需要满足多重约束和高性能的方向发展。模型预测控制(Modelpredictive control,MPC)得益于其在处理约束优化问题上的能力,近年来已在航天器控制中得到了广泛研究。本文首先回顾了MPC的发展历程,并基于该方法的基本形式,对其在航天器中的应用做了分类研究。然后,针对空间环境中的不确定性,依次总结鲁棒和随机两种MPC方法以及它们在航天器上的应用情况。在此基础上,考虑到航天器的安全性问题,总结现有研究中关于避障和容错的预测控制方法。针对MPC方法计算量大的缺陷,本文也从控制器的求解策略和求解方式出发,梳理了具有高计算效率的模型预测控制方法,包括分散式/分布式方法以及显式方法。最后,根据航天器的当前实际需求和未来发展,为MPC方法在航天器控制中的进一步研究提出了几点建议。
With increasing human space activities,space missions are evolving towards the need to meet multiple constraints and high performance.Due to the capability in handling constrained optimization problems,the model predictive control(MPC)has been widely studied in spacecraft control in recent years.This paper first reviews the development of MPC and categorizes its applications in spacecraft based on its basic form.Then,considering the uncertainty in space environments,the robust MPC and the stochastic MPC are summarized together with their applications.After that,the safety issues of spacecraft are taken into account and studies on obstacle avoidance and fault-tolerance are discussed in MPC framework.In addition,considering the high computational requirements of MPC,this paper also outlines some efficient methods from the perspectives of control strategy and optimization form,including decentralized/distributed MPC and explicit MPC methods.Finally,in consideration of the actual needs and future development of spacecraft,this paper provides several suggestions for further research on MPC in the spacecraft control aera.
作者
沈强
赵泽阳
康泽禹
张祎
SHEN Qiang;ZHAO Zeyang;KANG Zeyu;ZHANG Yi(School of Aeronautics and Astronautics,Shanghai Jiao Tong University,Shanghai 200240,China)
出处
《南京航空航天大学学报》
CAS
CSCD
北大核心
2023年第6期941-955,共15页
Journal of Nanjing University of Aeronautics & Astronautics
基金
国家自然科学基金青年科学基金(62103275)
国家自然科学基金联合基金(U20B2054,U20B2056)
国家重点研发计划(2020YFC2200800)。
关键词
航天器控制
模型预测控制
不确定性
安全性
计算效率
spacecraft control
model predictive control(MPC)
uncertainty
safety
computational efficiency
