Research on the inverse kinematics solution of robot arm for watermelon picking

Abstract

This paper puts forward a genetic algorithm based on non-linear programming in order to deal with the inverse kinematics solution precision of robot arm for watermelon picking, ensure the yield after picking and improve the fruit quality after picking. The robot arm for watermelon picking adopts the model of Denavit-Hartenberg, which mainly applies the non-linear genetic algorithm to give solution on the inverse kinematics issues. Lastly the paper differentiate the obtained inverse kinematics parameters through random forests algorithm. This paper respectively applies the genetic algorithm and nonlinear programming genetic algorithm for inverse kinematics solution on robot arm for watermelon picking with five degrees, six degrees and seven degrees of freedom. The experiment result shows that the non-linear programming genetic algorithm could effectively give inverse kinematics solution on robot arm for watermelon picking with multiple degrees of freedom, with the solution precision 300 to 600 times that of the genetic algorithm. There are more than one solution for the inversion result of the robot arm and the random forests algorithm could select fairly good picking path and gesture to reduce the unnecessary damage to the watermelon fruits in the picking.