Human-robot interaction through kinect and graphics tablet sensing devices

This paper analyzes three different solutions for human-robot cooperation. The proposed applications are based on two sensing devices: the Microsoft Kinect, a 3D camera by Microsoft, and a Graphics Tablet. The first solution, based on the Kinect, uses hand motion detection based on a clustering hand detection algorithm. The coordinates of the hands are sent to an application that computes the inverse kinematics and the desired robot trajectory. The second solution is based on a Graphic Tablet, which uses pen or finger input and operates in a similar way as the first system, but with higher accuracy. The third solution is also based on the Kinect but it uses the detection of the human body and voice commands together, to safely move the robotic arm without the need of a traditional teach pendant. All the three solutions adopt network sockets to send the commands to the robot controller. Behind the solutions, real time security procedures, monitors and avoids any collision of the robotic arm with humans during the movements.