摘要
提出了一种茄子收获机器人机械臂在笛卡尔空间的避障方法。将空间障碍物等效为可以用数学建模的圆柱扇环,将三维空间的路径规划简化为二维,提高了控制的实时性;将障碍从工作空间转换到C-空间中,使对机器人的控制直接作用于关节,避免了使用雅可比逆阵进行复杂的坐标转换。将C-空间映射到图像矩阵中,通过对图像进行适当的处理,规避了在使用A*算法寻优时可能出现的失败。实验结果表明,该避障路径规划方法计算量小,实时性好,适合自然生长状态下茄子的自动收获。
A method was put forward to avoid obstacles for an eggplant harvesting robot manipulator in Cartesian space. The obstacles were supposed equivalently as cylindrical-rings, and then a three- dimensional planning path problem was simplified to a two-dimensional one, which improved the realtime control performance greatly. Transforming the obstacles from work-space to configuration-space could control the robot joints directly, which could avoid complex coordinate translation using Jacobian inverse matrix. Moreover, processing properly image matrix mapped from C-space could avoid the path planning failure under A-star algorithm. The experimental results show that this algorithm with less computational amount and excellent real-time performance is suitable for harvesting natural growing eggplant automatically.
出处
《农业机械学报》
EI
CAS
CSCD
北大核心
2008年第11期94-98,共5页
Transactions of the Chinese Society for Agricultural Machinery
关键词
茄子收获机器人
机械手
避障
路径规划
Eggplant harvesting robot, Manipulator, Obstacle avoidance, Path planning
