-
摘要: 传统的奇异回避方法运算量大, 本文提出了一种新的 PUMA 类型机器人奇异回避方法—奇异分离加阻尼倒数法. 首先, 分析产生奇异的条件, 将导致 Jacobian 奇异的参数分离出来, 然后用阻尼倒数代替其普通倒数, 以回避运动学奇异的影响. 该方法无需对 Jacobian 进行 SVD 分解, 也无需估计其最小奇异值, 因而运算量小, 实时性好, 仅牺牲末端部分方向的精度, 适合于预定轨迹和实时轨迹的跟踪. 仿真和实验结果证明了算法的有效性.Abstract: Existing methods to avoid the kinematic singularities are complex. A novel approach (named "singularity separation plus damped reciprocal" method) is proposed for PUMA-type manipulators. Firstly, we analyze the singularity conditions and separate the singularity parameters from the Jacobian inverse. Then, the damped reciprocals are used to avoid the singularities by replacing their reciprocals. Since the SVD decomposition and the estimate of the minimum singularity value are not required, the algorithm is more efficient and fit for real-time application. And not all components of the end-effector velocities are sacrificed in accuracy. The method can be used for prescribed or real-time trajectory tracking. Simulation and experiment results verify the method.
计量
- 文章访问数: 3481
- HTML全文浏览量: 85
- PDF下载量: 2030
- 被引次数: 0