中文摘要：

提出一种工业机器人的最优轨迹规划方法。将机器人的轨迹视为由机器人关节空间中一系列的关键点构成,关键点两两之间通过三次多项式曲线进行连接。通过使用加权系数法定义代价函数,从而使机器人运动过程中的总动作时间和消耗能量在某种程度上达到综合最优,同时考虑关节速度、加速度、二次加速度以及力或力矩等约束条件。在代价函数的设计中,采用一种新颖的罚函数排序形式来处理约束问题。提出基因环境双演化免疫克隆算法对所定义的代价函数进行优化。以上策略的采用,使算法具备一定的学习能力,增强算法的全局搜索能力,从而提高解的质量和算法效率。对斯坦福机器人的仿真结果表明了本文方法与现有方法相比,具有更高的搜索效率,能得到性能更良好的解。

英文摘要：

A technique for optimal trajectory planning of robot manipulators is presented. It consists of linking two points in the operational space while minimizing a cost function, taking into account dynamic equations of motion as well as bounds on joint velocities, accelerations, jerks and force/torques. The cost function is used as a weighted balance of traveling time and mechanical energy of the actuators. Also, a novel ranking technique for the penalty function is designed to deal with the constraints. Furthermore, the environment-gene evolutionary immune clonal algorithm （EGICA） is proposed to solve the optimization problem. The use of the above-mentioned strategy makes the algorithm have a certain learning ability and enhances its global searching ability, thus improving solution quality and algorithm efficiency. The algorithm is tested with Stanford robot in simulation, and the result shows that the presented method has higher search efficiency and can obtain better solution in comparison with the existing methods.