Balance With a Two-Legged Robot With Artificial Neural Network Support

Robots are usually designed by inspiration from living creatures in the nature. It is inspired by the behaviors of 4 legged animals in the nature and various characteristics are given to robots. Many animals with four legs have the ability to walk and even walk new-born. The physiological developments of these creatures are completed in the mother’s womb and soon after birth, bone and leg muscle groups are available. It is a well-known fact that living beings can stay in balance by standing up to their legs after birth and carry out their neurological signals to the muscles of the legs. In this study, we have studied the robot kinematics and dynamics of 2-legged robot which resembles a 4-legged robot in the light of this study, and the robotic ability of the robot to stand on the ground with an artificial neural network(ANN) support. For the dynamic motion, the planar model of the robot is considered and 7-axis motion equations are obtained. The aim of the robot is to be able to stand up from the ground and stay in balance and it is aimed to be realized with the use of an artificial neural network structure as learning result. Conventional PID control method has been used for control signals that should be applied to robot joints. System responses were obtained graphically and the results were evaluated