Hang Dong, Ming Cong, and Heping Chen


Trajectory planning; Minimum-time joint trajectory; cascade Genetic Algorithms; Computer simulation


In trajectory planning problem of a robot manipulator in joint space, improving the working efficiency with high stability is a crucial issue in robotics. However, under most circumstances the cycle time is restricted by human experience. To find the optimal cycle time, many researches have been done in recent years. However, their algorithms were all based on an important yet hard-to define prior, that is, a well estimated operation time for any desired trajectory. If the operation time is not well estimated by senior technicians in a new job, their works will fail to provide a solution. Our paper highlights on solving the trajectory problems without that prior, which can automatically obtain the execution time and find a near-optimal time joint trajectory for a robot manipulator based on Hybrid Genetic Algorithms(HGA). Besides, this method can keep velocity, acceleration and jerk for each joint in bound along the trajectory. Therefore, the manipulator can perform a rapid yet smooth operation. In this paper, we solve this optimization problem in two stages with different objectives: quick approach stage and accurate search stage. For each stage, a unique type of Genetic Algorithm is implemented and its searching parameters are evolving according to a specific rule. Experiments were made to demonstrate the effectiveness of the proposed method. The results show that the algorithm can converge to a near-optimal solution quickly. A global solution of the robot kinematic scheme is also provided. Finally, the parameters of the Genetic Algorithms and coefficients used in the specific rules are evaluated. This paper opens a door for offline robot trajectory planning to satisfy a minimum-time yet high smooth movement requirement, hence the proposed method will innovate the robot joint trajectory planning process.

Important Links:

Go Back