When human and collaborative robots share their workspace, they manipulate objects together, exchange forces and they might even be in direct physical contact. In this context, the safety of human coworkers becomes a crucial aspect that classical collision avoidance cannot guarantee. This thesis introduces new concepts to minimize the potential physical harm that a robot can cause in quasi-static manipulations, especially in clamping situations. Using anthropometric data, as well as a priori geometric knowledge about the environment, we propose a method to continuously monitor possible clamping situations, and a control scheme to reduce their hazards. Therefore, we elaborate several circumstances that must apply in clamping situations and demonstrate an algorithm that checks their fulfillment in real-time using a hierarchy of Bounding Volumes. Robot manipulation is rendered safe by limiting the contact forces adaptively in clamping situations.     «

When human and collaborative robots share their workspace, they manipulate objects together, exchange forces and they might even be in direct physical contact. In this context, the safety of human coworkers becomes a crucial aspect that classical collision avoidance cannot guarantee. This thesis introduces new concepts to minimize the potential physical harm that a robot can cause in quasi-static manipulations, especially in clamping situations. Using anthropometric data, as well as a priori...     »

When human and collaborative robots share their workspace, they manipulate objects together, exchange forces and they might even be in direct physical contact. In this context, the safety of human coworkers becomes a crucial aspect that classical collision avoidance cannot guarantee. This thesis introduces new concepts to minimize the potential physical harm that a cobot can cause in quasi-static manipulations, especially in clamping situations. Using anthropometric data, as well as a priori geometric knowledge about the environment, we propose a method to continuously monitor possible clamping situations, and a control scheme to reduce their hazards. Therefore, we elaborate several circumstances that must apply in clamping situations and demonstrate an algorithm that checks their fulfillment in real-time using a hierarchy of Bounding Volumes. Robot manipulation is rendered safe by limiting the contact forces adaptively in clamping situations.     «

When human and collaborative robots share their workspace, they manipulate objects together, exchange forces and they might even be in direct physical contact. In this context, the safety of human coworkers becomes a crucial aspect that classical collision avoidance cannot guarantee. This thesis introduces new concepts to minimize the potential physical harm that a cobot can cause in quasi-static manipulations, especially in clamping situations. Using anthropometric data, as well as a priori...     »