Grasp Planning Based on Metrics for Collaborative Tasks Using Optimization

In recent years, technological developments in the field of robotics have expanded their application spectrum to encompass tasks that involve human inclusion in the same workspace. One of the challenges of robotics collaboration is the issue of how a robot and a human can perform daily collaborative tasks, like manipulation of an object. One significant specific problem to solve is where the robot can grasp the object knowing the human grasping points. This research proposes a planning algorithm to find a robot grasping point based on geometric grasp metrics as well as a new heuristic metric focused on the intrinsic inertia in multi-directional object movement. We propose three grasping points: two points emulating each human hand, positioned anywhere on the object and one last point, referencing the robot, which will be optimized as a multi-objective (MO) function problem. The planner was tested using common objects present in human environments (a chair and a table).