摘要
基于黑板模型设计了排球机器人动作规划框架,并将动作规划中的各个子任务定义为智能体。由黑板激活相应的智能体进行子任务求解,从而完成排球机器人的击球动作规划问题。此外,考虑到排球任务的特点,以机械臂的可操作度作为评价函数,选取最优的机械臂构形。最后,采用MatlabSimulink对平面3连杆机械臂的击球问题进行了仿真研究,得到满足可操作度指标下的最优击球动作。该仿真的实现过程体现了基于黑板模型及多智能体技术的思想。
We design a motion planning frame for the volleyball robot task based on blackboard model, and refer each subtask as an agent. The problem of motion planning is settled by the solution of each subtask activated by blackboard. Besides, the optimal posture of the manipulator in its feasible motion set is chosen by computing the manipulability measure of the manipulator, which differs from the motion planning of other ball-playing robots. The simulation of a 3-link planar manipulator hitting ball task is implemented using Matlab Simulink and obtains the optimal hitting motion subject to manipulability measure. The simulation also illustrates the idea of motion planning based on blackboard and multi-agent technique.
出处
《机械科学与技术》
CSCD
北大核心
2005年第5期581-583,630,共4页
Mechanical Science and Technology for Aerospace Engineering
