Programming and neural-net stabilizing of manipulator timeoptimal trajectories with construction constraints

The problem of programming and neural-net stabilizing of manipulator time-optimal trajectories is taken into consideration within context of neural-nets threatening for inverse dynamic model approximation. Trajectories are put into phase plane and generalized coordinates direct dependences on time and constraints are obtained. The problem of links movements’ synchronization is addressed in brief. A nonlinear control law with neural-net representation of inverse dynamic model is illustrated with simulations of 5 DOF manipulator control (without consideration of servo dynamics, i.e. Control signals are generalized forces). Method addresses time-optimal real-time control of multilinked nonlinear system.