Abstract
In this paper, an indirect method for trajectory planning for industrial robots has been addressed using an evolutionary algorithm. The algorithm is divided into three stages: (1) The acquisition of Adjacent Configurations (AC) for Path Planning subjected to kinematics, geometric and obstacle avoidance constraints. (2) The acquisition of a collision-free path between initial and goal robot configurations. This path consists of a set of ACs, and (3) The acquisition of a temporal history of the evolution for the robot joint coordinates, by minimizing the required time subjected to actuator limits. This algorithm has been evaluated by comparing the results with the direct procedures proposed by Rubio articles in 2009 and 2010.
1. Introduction
In the last few decades, the prevalence of robots has grown in many areas. In addition to industrial applications, robots are also used in surgery, agriculture, underwater, and for transportation. In industrial applications, they have many purposes like; pick and place operations, assembly tasks, spray-painting, and many other tasks.
In general, the purpose of robotic systems is to perform such tasks smoothly in as little time as possible. Thus, a procedure of three stages is proposed to obtain an offline minimum time trajectory planning. The offline trajectory planning is justified as a large number of robotic applications work in a repetitive manner.
7. Conclusion
In this paper, a new three-staged evolutionary algorithm for trajectory planning problem for industrial robots in connection with obstacle avoidance has been presented, in this case exemplarily shown for different workspace scenarios using a PUMA 560 six-axis manipulator. This method combines three formerly known evolutionary algorithms to solve the three stages, which are
(1) acquisition of possible and collision-free configurations using SSGA, (2) path planning with minimum path length under kinematic constraint and (3) minimum time trajectory planning conditioned by the robot’s dynamics. The main benefit of the presented method is its performance increase compared to common methods concerning execution time and traveling distance
