Path planning using potential field algorithms with optimized parameters applied to a 6-DOF anthropomorphic manipulator

• Autores: Alcántara Tacora, Sandro Manuel, López Zapata, Erwin Daniel, Peralta Toribio, Jesús, Rodríguez Bustinza, Ricardo Raúl
• Formato: artículo
• Fecha de Publicación: 2021
• Institución: Universidad Nacional de Ingeniería
• Repositorio: Revistas - Universidad Nacional de Ingeniería
• Lenguaje: español
• OAI Identifier: oai:oai:revistas.uni.edu.pe:article/848
• Enlace del recurso: https://revistas.uni.edu.pe/index.php/tecnia/article/view/848
• Nivel de acceso: acceso abierto
• Materia: Planificación de movimiento; campos potenciales; red neuronal supervisada; cinemática inversa; manipulador robótico; Path planning; Potential field; Supervised Neural Network; Inverse Kinematics; Robotic Manipulator

This paper develops a variation of potential field algorithm for obstacle avoidance trajectory planning applied to an anthropomorphic manipulator of 6 degrees of freedom. In the first instance, the inverse kinematics model was generated based on a multivariate iterative control process, then the model was modified by adding a rotation vector obtained by the repulsive forces between the obstacle and the six joints of the robot so that the manipulator can find a route that avoids the obstacle and reaches a goal position. This final model of inverse kinematics with potential fields generates trajectories that depend on the optimization step size parameter and the vector fields coefficients. In order to optimize the trajectories, a database was generated with the initial, final and obstacle points of different trajectories with their respective parameters optimized to train a supervised neural network. The results show that the neural network must be trained with a greater amount of data because it calculates erroneous parameters for certain initial and final positions. Finally, the simulation of the 6dof manipulator that follows the trajectory generated by the inverse kinematics model and potential fields with optimized parameters empirically calculated results in a model with an appropriate behavior managing to avoid obstacles.