This figure presents three distinct scenarios tested in the study. Scenario 1 (left) features three movable obstacles (white with fiducial markers) and one target object (yellow). Scenario 2 (center) introduces a tunnel obstacle partitioning the table. Scenario 3 (right) includes 11 movable obstacles and one target object. The green shaded area represents the goal for each Scenario.

This figure presents three distinct scenarios tested in the study. Scenario 1 (left) features three movable obstacles (white with fiducial markers) and one target object (yellow). Scenario 2 (center) introduces a tunnel obstacle partitioning the table. Scenario 3 (right) includes 11 movable obstacles and one target object. The green shaded area represents the goal for each Scenario.

Source publication
Preprint
Full-text available
This work explores non-prehensile manipulation (NPM) and whole-body interaction as strategies for enabling robotic manipulators to conduct manipulation tasks despite experiencing locked multi-joint (LMJ) failures. LMJs are critical system faults where two or more joints become inoperable; they impose constraints on the robot's configuration and con...

Contexts in source publication

Context 1
... 2 (Fig. 3, right): In the second case, two joints are locked: the wrist lift and wrist rotation joints. This failure case precludes the use of the end-effector in completing manipulation tasks. This enables us to test the system using exclusively the non-end-effector parts of the robot's embodiment. The experimental scenarios are presented in Fig. 5. Each scenario involves manipulating the target object (the yellow cube) to the goal location (green ...
Context 2
... Scenario 1 (Fig. 5, left): This scenario has three movable obstacles and one target object. This scenario is motivated by the need to be able to manipulate in cluttered environments that hinder direct access to the target ...
Context 3
... Scenario 2 (Fig. 5, center): The second scenario has a tunnel obstacle in the middle, partitioning the two halves of the table. This scenario is motivated by the need to perform dexterous manipulation while maneuvering around environmental ...
Context 4
... Scenario 3 (Fig. 5, right): There are 11 movable obstacles and one target object. This scenario allows us to explore the utility of our planner in the presence of increased object-obstacle interactions. To comprehensively understand how NPM abilities expand the robot's task space, we conduct the reachability analysis in Section V-A. In Section V-B, we show the ...
Context 5
... This scenario allows us to explore the utility of our planner in the presence of increased object-obstacle interactions. To comprehensively understand how NPM abilities expand the robot's task space, we conduct the reachability analysis in Section V-A. In Section V-B, we show the effectiveness of our system across the three scenarios shown in Fig. 5, for the two failure cases described in Fig. 3, using the three Action Selection Mechanisms described in Section III-D over a total of 267 trials. This analysis shows the limitations of grasping-only actions and demonstrates the improvements brought about by NPM abilities. Through these results, we seek to verify two hypotheses: H.1 ...

Similar publications

Conference Paper
Full-text available
This work explores non-prehensile manipulation(NPM) and whole-body interaction as strategies for enabling robotic manipulators to conduct manipulation tasks despite experiencing locked multi-joint (LMJ) failures. LMJs are critical system faults where two or more joints become inoperable; they impose constraints on the robot’s configuration and cont...