ETD

• elastic actuators
• nonlinear optimal control
• optimal control
• optimization
• robotics

We presented a method for obtaining numerical solutions of optimal control
problems (OCP) for systems of ordinary differential equations with possible appli-
cations in dynamic tasks in combination with predictive control algorithms based
on the model. The method is intended for applications that require the resolution
of OCP in a short time and can efficiently exploit the prior knowledge ( obtained
for example by the application of the Principle of Maximum) of control’s arcs by
determining their length and sequence.
The presented method has been applied in simulation and to a real prototype
of planar robotic manipulator with two degrees of freedom with elastic joints.
Experimentally it was obtained that the system in the tested configuration is
able to reach a speed higher than double that attainable without elastic joints .
The approach with predictive control enables a reduction of the error on the desired
final pose of the robot up to 10 times compared to about an experiment conducted
with control in open loop.