柔性空间机器人姿态运动规划的混合优化策略

发布时间：2018-07-22 15:06

【摘要】：自由漂浮是空间机器人执行任务时常用的工作模式,其姿态规划是完成复杂空间任务的基础.针对自由漂浮柔性空间机器人姿态运动规划问题,提出一种由Gauss伪谱法求解可行解与直接打靶法求解更为精确的数值解相结合的混合优化策略.首先,利用假设模态法近似描述柔性臂杆的弹性变形,根据Lagrange方程建立空间机器人系统的动力学模型;然后,利用Gauss伪谱法将空间机器人非完整运动规划问题离散为非线性规划问题,求解在较少Legendre-Gauss(LG)点时状态变量和控制变量对应的可行解,可行解的求解采用从可行解到近似最优解的串行优化策略;最后,在LG点处离散控制变量,作为直接打靶法的初值,利用序列二次规划算法求解最优的控制输入,再通过数值积分得到空间机器人系统姿态运动优化轨迹.通过数值仿真,求解得到的姿态运动曲线光滑平稳,最优控制输入也能很好地满足各种约束条件.结果表明该混合优化策略具有很好的鲁棒性和有效性.
[Abstract]:Free floating is a common mode of work for space robots, and its attitude planning is the basis for accomplishing complex space tasks. A hybrid optimization strategy combining Gauss pseudospectral method with direct shooting method to solve the problem of attitude motion planning of flexible free-floating space robot is proposed. Firstly, the elastic deformation of the flexible arm rod is approximately described by the assumed modal method, and the dynamic model of the space robot system is established according to the Lagrange equation. The nonholonomic motion planning problem of space robot is discretized into a nonlinear programming problem by Gauss pseudospectral method, and the feasible solutions of state variables and control variables are solved when there are fewer Legendre-Gauss (LG) points. The feasible solution is solved by serial optimization strategy from feasible solution to approximate optimal solution. Finally, as the initial value of direct shooting method, the discrete control variable is discretized at LG point, and the optimal control input is solved by using sequential quadratic programming algorithm. Then the optimal trajectory of attitude motion of space robot system is obtained by numerical integration. Through numerical simulation, the obtained attitude motion curve is smooth and stable, and the optimal control input can satisfy all kinds of constraint conditions well. The results show that the hybrid optimization strategy is robust and effective.
【作者单位】：北京信息科技大学机电工程学院;北京信息科技大学理学院;
【基金】：国家自然科学基金(11472058)
【分类号】：TP242

【相似文献】