The basic experimental paradigm we use is the reaching towards a cylindrical object with the goal to grasp it with a precision side grip at a fixed height. Variations of this basic paradigm allowed us to specifically test different aspects of the planning model.
The design of some of the scenarios presented in this chapter arose from the collaboration with the NICI - Nijmegen Institute for Cognition and Information (as of September 1st, 2008, NICI is part of the Donders Institute for Brain, Cognition and Behavior at Radboud University, Nijmegen, The Netherlands).
An experimental study that focused on prehension kinematics that typically reflect, collision-avoidance characteristics of grasping movements, was conducted at NICI. The results were published in (Lommertzen et al., 2008). Some early results of movement planning were presented in (Costa e Silva et al., 2009). The results presented in this chapter can be found in (Costa e Silva, under review).
Reaching and grasping a tall target object
Reaching and grasping a tall target object in the
presence of an additional obstacle
ARoS grasps a tall cylindrical target object (a packet of biscuits), to hand it over to the human, by moving the hand around the object.
Reaching and grasping a shallow target object
ARoS reaching to grasp a tall object (a packet of biscuits) to hand it over to the human in the presence of an intermediate shallow obstacle. ARoS avoids the collision by passing above the intermediate obstacle.
The robot reaches and grasps the shallow target object (tea-caddy). Here, the collision with the object to be grasped is avoided by passing above it.
When the intermediate obstacle is tall, ARoS avoids the collision by going around it.
For the same scenario as in Figure 8.9, the behavior of the robot is different, since the collision with the object to be grasped is avoided by going around it.