Chapter 10 – Real-time performance of the model

In this chapter we evaluate the real-time performance of the model proposed in this thesis for generating trajectories of a high degree of freedom (DOF) robotics arm-hand system that reflects optimality principles of human motor control and key characteristics of human arm trajectories (Todorov, 2004).

We have selected three solvers that are adequate for the purpose of this study, namely, IPOPT (wachter and Biegler, 2007), KNITRO (Bryrd et al., 2006) and SNOPT (Gill et al., 2002), since they are all well regarded and recognized solvers for large-scale nonlinear optimization.

The results presented in this chapter have been published in (Costa e Silva et al., 2011).

The performance of the three nonlinear optimization solvers has been tested in four different problems (see Table 10.1), in which the anthropomorphic robot ARoS has to grasp different objects,with different grip types, thereby avoiding several obstacles. 

Specifically, we focus here on reaching and grasping columns and wheels, using a side and an above grip, respectively (see Figure 10.1).

See the subpages for the videos of each Problem.