Human Productivity in a Workspace Shared with a Safe Robotic Manipulator

Abstract

Robotics technology is approaching a state where highly capable autonomous platforms can directly interact with human companions. This paper explores the ability of a human and manipulator to efficiently complete independent tasks in a shared physical workspace without explicit communication of their planned actions. With focus on quantifying the impact of robot presence on human task performance and workload, a series of human-subject experiments were conducted in which a seated human was asked to perform cognitive and mobility tasks in this shared workspace. The manipulator could unintentionally interrupt these activities, blocking the field of view or physical path to a worksite while completing its own tasks. To avert impending conflict, the robot received and reacted to ideal sensor data identifying the human’s tasks upon initiation. Tests over a progressively complex set of tasks for the human–manipulator team were conducted. Results indicate minimal impact on a human companion’s productivity when a manipulator is introduced to the workspace, provided it responds quickly to avoid blocking constraints. Future work involves maturation of the robot’s perception and prediction capabilities followed by additional shared workspace experiments.