Abstract: A comprehensive optimal trajectory planning method was proposed to improve the efficiency and impact resistance of robotic arms. Firstly, by establishing a 3-5-3 polynomial curve mathematical model, a joint motion trajectory with controllable endpoint motion parameters was constructed. Secondly, considering constraints such as joint position, velocity, and acceleration, the objective function was defined using the weighted coefficient method to achieve a comprehensive optimization of the action time, impact, and dexterity of the robotic arm. The dynamic weighting method was used in the design of the objective function to address the contradiction between joint velocity and impact. Finally, for the standard particle swarm optimization algorithm, the Latin hypercube sampling function was used to homogenize the population, and a random inertia weight update strategy was proposed to obtain an improved particle swarm algorithm. This algorithm was used to optimize the objective function and obtain the comprehensive optimal motion trajectory. Simulation and prototype experiments were conducted, and the experimental results demonstrate the feasibility of the proposed method.

Key words: industrial robot, trajectory planning, multi objective optimization, particle swarm optimization(PSO)

摘要： 为提高机械臂作业效率以及抗冲击能力，提出了一种综合最优轨迹规划方法。首先通过建立3-5-3多项式曲线数学模型构造出端点运动参数可控的关节运动轨迹；然后考虑关节位置、速度、加速度等约束条件，通过加权系数法定义目标函数，使机械臂的动作时间、冲击和灵巧度达到综合最优，在目标函数的设计中采用动态加权方法来处理关节速度与冲击之间的矛盾；最后，针对标准粒子群算法，利用拉丁超立方抽样函数均匀化种群，并提出随机惯性权重更新策略，得到改进粒子群算法，利用该算法对目标函数进行优化，得到综合最优运动轨迹。进行了仿真和样机实验，实验结果证明所提方法具有可行性。

关键词: 工业机器人, 轨迹规划, 多目标优化, 粒子群算法