CONTROL OF A REDUNDANT ARTICULATED SYSTEM BY NEURAL NETWORKS

Neural networks are used to solve a nonlinear control problem in robotics. A neural network has learned how to control the angles of the front and back forks of a simplified mobile robot (a bicycle) which moves forward on two-dimensional uneven terrain in order to maintain its higher bar horizontal. A poor knowledge of the model is used. The architecture of the neural network depends only on the number of variables and the number of degrees of freedom. Although no mathematical proof exists, computer simulation has demonstrated workability. The ground not being known in advance, the bicycle adapts itself to it as it moves forward.