基于PSO-SA的机器人关节空间轨迹规划

TRAJECTORY PLANNING OF ROBOT JOINT SPACE BASED ON PSO-SA

针对工业机器人点到点轨迹规划问题,提出一种基于PSO-SA的时间最优机器人关节空间轨迹规划方法。使用模拟退火算法(SA)对粒子群算法(PSO)进行优化,将模拟退火机制引入到粒子群算法以提高算法的全局搜索能力。使用惯性权重非线性递减策略以及动态学习因子来平衡算法的全局与局部搜索能力。以PUMA_560机器人作为研究对象,通过5-7-5多项式插补函数得到各关节的轨迹曲线。通过PSO-SA优化关节运动时间,并加入关节的速度和加速度约束。对前三个关节进行实验仿真,结果表明PSO-SA比传统的PSO能得到更短的轨迹时间,算法也有更好的稳定性,提高了机器人的运动效率。

Aiming at the problem of point-to-point trajectory planning for industrial robots, we propose a time-optimal trajectory planning of robot joint space based on PSO-SA. The simulated annealing(SA) algorithm was used to optimize the particle swarm optimization(PSO). We introduced SA into PSO-SA to improve the global search capability. We adopted inertial weight non-linear decreasing strategy and dynamic learning factor to balance the global and local search capabilities. Taking PUMA_560 robot as the research object, we obtained the trajectory curve of each joint through the 5-7-5 polynomial interpolation function. The joint motion time was optimized by PSO-SA, and the joint speed and acceleration constraints were added as constrained condition. The experimental simulation of the first three joints shows that PSO-SA can obtain a shorter trajectory time than the traditional PSO. This algorithm has better stability and improves the robot’s movement efficiency.