Trajectory Planning of Multijointed Link Structures Using a Neural Network
Junji Tani, Hiroki USUDA · 1993
Intelligent/adaptive structures can change their geometrical configuration and mechanical characteristics by themselves. In order to construct large space structures, it is important to plan how to operate them effectively. In this paper, a neural network is used in order to plan effective operation for a multijointed link structure which is a simplified model of intelligent/adaptive structures. The task considered here is to generate an effective joint trajectory for development of a multijointed link structure from the packaged state to the final configuration. Numerical simulations and experiments show that a neural network can solve such a problem under minimum torque.