基于随机森林的月球表面软着陆实时最优控制
|
摘要
针对传统月球表面软着陆在处理入轨偏差或降落过程轨迹偏离实时性差,文章提出一种通过监督学习离线训练随机森林结构,使得着陆器在降落过程中根据其状态,通过训练的随机森林结构在线计算其控制量,从而达到实时控制的目的。文章还提出一种基于随机森林的模型对月球着陆过程轨迹重规划技术,通过动力学建模将月球着陆过程分成制动段、接近段和着陆段共3个阶段,利用离线训练好的模型根据航天器状态在线计算其控制量,并通过三段下降过程逐渐降低航天器位置速度误差。仿真结果表明,针对入轨偏离500m的情况,通过第一阶段将位置误差缩短至50m,保证了航天器状态位于第二阶段训练集内,经过接近段后再将位置误差缩小至10m范围内,速度误差降至0.01m/s量级,满足着陆误差要求,且控制量计算时间短,达到了轨迹实时重规划的目的。
Traditional optimal trajectories for lunar soft landing can′t be calculated in real time if there′s error when spacecraft entering orbit or during the descending stage.A novel scheme was proposed to replan the trajectory in real time for the random forest model.The descending stage was modeled by three phases:deboost,descend and land.The model can predict the control by the state of spacecraft as the input,the lander descends lower and the error gets smaller.The simulation results show that with a 500 m error in entering orbit the distances deceases to 50 m after the deboost phase.This error is in the range of the trainingset of descend phases,so after that the position error is less than 10 m and the velocity error is less than 0.01 m/s,both in tolerance.Meanwhile this method is fast enough to satisfy the real-time control.
Traditional optimal trajectories for lunar soft landing can′t be calculated in real time if there′s error when spacecraft entering orbit or during the descending stage.A novel scheme was proposed to replan the trajectory in real time for the random forest model.The descending stage was modeled by three phases:deboost,descend and land.The model can predict the control by the state of spacecraft as the input,the lander descends lower and the error gets smaller.The simulation results show that with a 500 m error in entering orbit the distances deceases to 50 m after the deboost phase.This error is in the range of the trainingset of descend phases,so after that the position error is less than 10 m and the velocity error is less than 0.01 m/s,both in tolerance.Meanwhile this method is fast enough to satisfy the real-time control.
引文
[1]彭祺擘,李海阳,沈红新,等.基于Gauss伪谱法和直接打靶法结合的月球定点着陆轨道优化[J].国防科技大学学报,2012,34(2):119-124.PENG Q B,LI H Y,SHEN H X,et al.Lunar exactlanding trajectory optimization via the method combining GPM with direct shooting method[J].Journal of National University of Defense Technology,2012,34(2):119-124(in Chinese).
[2]胡海龙,南英,闻新.基于遗传算法的航天器再入动态终迹圈研究[J].中国空间科学技术,2014,34(1):35-41.HU H L,NAN Y,WEN X.Research on dynamic reentry footprint for the spacecraft based on the genetic algorithm[J].Chinese Space Science and Technology,2014,34(1):35-41(in Chinese).
[3]梁栋,刘良栋,何英姿.月球精确软着陆最优标称轨迹在轨制导方法[J].中国空间科学技术,2011,31(6):27-35.LIANG D,LIU L D,HE Y Z.On-board optimal nominal trajectory guidance method for lunar pinpoint soft landing[J].Chinese Space Science and Technology,2011,31(6):27-35(in Chinese).
[4]梁栋,刘良栋,何英姿.月球精确软着陆李雅普诺夫稳定制导律[J].中国空间科学技术,2011,31(2):25-31.LIANG D,LIU L D,HE Y Z.Guidance law for lunar pinpoint soft landing based on Lyapunov stability theory[J].Chinese Space Science and Technology,2011,31(2):25-31(in Chinese).
[5]ACIKMESE B,PLOEN S R.Convex programming approach to powered descent guidance for mars landing[J].Journal of Guidance,Control,and Dynamics,2007,30(5):1353-1366.
[6]HUANG B Q,CAO G Y,GUO M.Reinforcement learning neural network to the problem of autonomous mobile robot obstacle avoidance[C].International Conference on MachineLearning and Cybernetics,Guangzhou,China,18-21August,2005,IEEE:85-89.
[7]WIDMAIER F.Robot arm tracking with random decision forests[M].Tubingen:Eberhard-Karis University Tubingen,2015:18-22.
[8]LI Q,QIAN J,ZHU Z,et al.Deep neural networks for improved,impromptu trajectory tracking of quadrotors[C].International Conference on Robotics and Automation(ICRA),Singapore,29 May-3June,2017,IEEE:5183-5189.
[9]王晓晖,李爽.考虑动态不确定因素的深空探测器任务规划[J].中国空间科学技术,2016,36(6):29-37.WANG X H,LI S.A mission planning method for deep space explorer considering dynamic uncertainties[J].Chinese Space Science and Technology,2016,36(6):29-37(in Chinese).
[10]U.S.trademarkregistration number 3185828,registered 2006/12/19.
[11]PIERSON B L,KLAEVER C A.Three-stage approach to optimal low-thrust earth-moon trajectories[J].Journal of Guidance,Control,and Dynamics,1994,17(6):1275-1282.
[12]IZZO D,DE CROON G.Nonlinear model predictive control applied to vision-based spacecraft landing[C].Conference on Guidance,Navigation and Control,Delft,Netherlands,10-12 April,2013:91-107.
[13]SNCHEZ-SNCHEZ C,IZZO D,HENNES D.Learning the optimal state-feedback using deep networks[C].Symposium Series on Computational Intelligence(SSCI),Athens,Greece,5-9 December,2016,IEEE:1-8.
[14]RAO A V,BENSON D A,DARBY C,PATTERSON M A,et al.Algorithm 902:GPOPS,a MATLAB software for solving multiple-phase optimal control problems using the Gauss pseudospectral method[J].ACM Transactions on Mathematical Software,2010,37(2):22.
[15]GILL P E,MURRAY W,SAUNDERS M A.SNOPT:an SQP algorithm for largescale constrained optimization[J].SIAM Review,2005,47(1):99-131.
[16]ZHOU Z H.Ensemble methods:foundations and algorithms[M].Boca Raton:CRC press,2012.
[17]SNCHEZ-SNCHEZ C,IZZO D,HENNES D.Optimal real-time landing using deep networks[C]∥Proceedings of the Sixth International Conference on Astrodynamics Tools and Techniques,Darmstadt,Germany,14-17 March,2016,ICATT:2493-2537.
[2]胡海龙,南英,闻新.基于遗传算法的航天器再入动态终迹圈研究[J].中国空间科学技术,2014,34(1):35-41.HU H L,NAN Y,WEN X.Research on dynamic reentry footprint for the spacecraft based on the genetic algorithm[J].Chinese Space Science and Technology,2014,34(1):35-41(in Chinese).
[3]梁栋,刘良栋,何英姿.月球精确软着陆最优标称轨迹在轨制导方法[J].中国空间科学技术,2011,31(6):27-35.LIANG D,LIU L D,HE Y Z.On-board optimal nominal trajectory guidance method for lunar pinpoint soft landing[J].Chinese Space Science and Technology,2011,31(6):27-35(in Chinese).
[4]梁栋,刘良栋,何英姿.月球精确软着陆李雅普诺夫稳定制导律[J].中国空间科学技术,2011,31(2):25-31.LIANG D,LIU L D,HE Y Z.Guidance law for lunar pinpoint soft landing based on Lyapunov stability theory[J].Chinese Space Science and Technology,2011,31(2):25-31(in Chinese).
[5]ACIKMESE B,PLOEN S R.Convex programming approach to powered descent guidance for mars landing[J].Journal of Guidance,Control,and Dynamics,2007,30(5):1353-1366.
[6]HUANG B Q,CAO G Y,GUO M.Reinforcement learning neural network to the problem of autonomous mobile robot obstacle avoidance[C].International Conference on MachineLearning and Cybernetics,Guangzhou,China,18-21August,2005,IEEE:85-89.
[7]WIDMAIER F.Robot arm tracking with random decision forests[M].Tubingen:Eberhard-Karis University Tubingen,2015:18-22.
[8]LI Q,QIAN J,ZHU Z,et al.Deep neural networks for improved,impromptu trajectory tracking of quadrotors[C].International Conference on Robotics and Automation(ICRA),Singapore,29 May-3June,2017,IEEE:5183-5189.
[9]王晓晖,李爽.考虑动态不确定因素的深空探测器任务规划[J].中国空间科学技术,2016,36(6):29-37.WANG X H,LI S.A mission planning method for deep space explorer considering dynamic uncertainties[J].Chinese Space Science and Technology,2016,36(6):29-37(in Chinese).
[10]U.S.trademarkregistration number 3185828,registered 2006/12/19.
[11]PIERSON B L,KLAEVER C A.Three-stage approach to optimal low-thrust earth-moon trajectories[J].Journal of Guidance,Control,and Dynamics,1994,17(6):1275-1282.
[12]IZZO D,DE CROON G.Nonlinear model predictive control applied to vision-based spacecraft landing[C].Conference on Guidance,Navigation and Control,Delft,Netherlands,10-12 April,2013:91-107.
[13]SNCHEZ-SNCHEZ C,IZZO D,HENNES D.Learning the optimal state-feedback using deep networks[C].Symposium Series on Computational Intelligence(SSCI),Athens,Greece,5-9 December,2016,IEEE:1-8.
[14]RAO A V,BENSON D A,DARBY C,PATTERSON M A,et al.Algorithm 902:GPOPS,a MATLAB software for solving multiple-phase optimal control problems using the Gauss pseudospectral method[J].ACM Transactions on Mathematical Software,2010,37(2):22.
[15]GILL P E,MURRAY W,SAUNDERS M A.SNOPT:an SQP algorithm for largescale constrained optimization[J].SIAM Review,2005,47(1):99-131.
[16]ZHOU Z H.Ensemble methods:foundations and algorithms[M].Boca Raton:CRC press,2012.
[17]SNCHEZ-SNCHEZ C,IZZO D,HENNES D.Optimal real-time landing using deep networks[C]∥Proceedings of the Sixth International Conference on Astrodynamics Tools and Techniques,Darmstadt,Germany,14-17 March,2016,ICATT:2493-2537.