基于MDD不动点迭代的控制分配方法

Control allocation method based on MDD fixed-point iteration

针对航空航天领域中存在的控制分配问题,提出一种基于最大方向导数(MDD)不动点迭代的控制分配策略.定义MDD为当前迭代点沿所有单位向量方向中最大导数对应的单位向量,沿MDD方向,取前一迭代点和当前迭代点的迭代增量作最小二乘拟合,下一迭代点在该方向上的估计迭代增量作为当前迭代点的更新迭代增量,并对改进算法的收敛性进行证明.通过不同方法的仿真对比验证,表明了所提出算法的有效性和快速性.

@@@@For the control allocation problem in the field of aeronautics and astronautics, a control allocation method based on max direction derivative(MDD) fixed-point(FXP) iteration is proposed. MDD is defined as the unit vector which has the max derivative among all the directions of the unit vector at current iteration point. Along the MDD direction, the increment of next iteration point is estimated by using least square fitting of the increments of the former and current iteration point, which is taken as a update for the increment of current iteration along the MDD direction. The convergence of the method is also proved. Finally, the comparison of different methods is carried out. The simulation results show the effectiveness and the rapidity of the proposed method.