Search for: All records

Interfacing between robots and humans has come a long way in the past few years, and new methods for smart, robust interaction are needed. Typically, a technician has to program a routine for a robot in order for the robot to be useful. This puts up a significant barrier to entry into the field of automating tasks using robots—not only is a technician and a computer required, but the robot is not adaptive to the immediate needs of the user. The robot is only capable of executing a pre-determined task and for any change to be made the entire system needs to be paused. This project seeks to bridge the gap between user and robot interface, creating an easy-to-use system that allows for adaptive robot control. Using a combination of computer vision and a monocular camera system and integrated LiDAR sensor on an iPhone, gesture recognition and pose estimation was conducted within an independent system to control the Baxter humanoid robot. The gathered data was sent wirelessly to the robot to be interpreted and then replay actions performed by the user.