TeleSense: A Teleoperation System by Using Embedded Proximity Sensors and Leap Motio

In this thesis we present a new teleoperation system which is designed to let a user controls a robot easier. In many teleoperation systems which are currently used, the master controls the slave robot through a joystick while observing its behavior by the help of a remote camera on a screen. Controlling a robot in this situation is not an easy task and needs trained users to do it. The first problem is that moving a robot's end effector by using a joystick in a 3D space is not intuitive. The second problem is that a 2D camera doesn't transfer depth information to the master. In such systems perceiving objects depth for a human is difficult. The lack of ability to perceive depth in such scenarios can be problematic during many tasks that collision with objects might be undesirable such as moving a glass full of water in daily activities or removing a tumor in a telesurgery. As robots become more pervasive and are used by a vast majority of people a more comfortable system for controlling them remotely is desired. This work presents such a system in which for the sake of user comfortability we replace the joystick controller with a Leap Motion device which directly maps the movements of a human's hand to the robot's end effector with an arbitrary scale and arbitrary precision. Moreover, in an attempt to find a solution for the second problem we augmented the robot's end effector with proximity sensors which let the user find out how close an object is to the robot's fingers without physical contact. We demonstrated the effectiveness of our teleoperation system by conducting a user-study in which 4 humans participate in a pick and place experiment in different settings. We observed that embedded proximity sensors can improve user experience a lot and resulted in 31% speedup in average in terms of task completion time.