Application of Genetic Algorithm for Trajectory Planning of Two Degrees of Freedom Robot Arm With Dimensions

Abstract

In this work, the trajectory planning problem of two-link robot arm in the two dimensions workspace with containing obstacles was considered. The objectives of this research were to apply genetic algorithm (GA) to minimise the rotation angles of two-joint robot arm. The algorithm was used to find interior points, which were then used to interpolate the robot trajectory by avoiding obstacles. The methods for point-to-point interpolation including linear and cubic spline interpolation methods were investigated. However, due to the nature of stochastic search conducted within the GA, the investigation of this study was not only focused on the quality of the solution itself but also on the investigation of appropriate setting of GA parameters. Haft fractional factorial experimental design was carried out with five replications. The results were then analyzed using a general linear form of analysis of variance and main effect plots. It was found that the experimental results obtained by using the cubic spline interpolation method generally produced better results than those of using linear spline interpolation method. The analysis of variance on the results suggested some GA parameters were found to be statistically significant factors.