Pose Detection-Based Real-Time Control of Robotic Manipulator for Remote Locations

In this paper, we have developed a pose detection-based control of a serial manipulator to operate in remote locations, which uses a pose detection deep learning algorithm for pose estimation. The presented control strategy for a serial manipulator has opened many doors for a variety of applications in various sectors like the teleoperation of humanoid robots, VR games, etc. With pose detection or pose detection-based control, we need very less computational power to find the pose of the user. We are controlling the robot by task space mapping of human actions based on the robot’s environment to the robot’s actions in the manipulator’s coordinate frame in an untrained robot environment and using Jacobian-based inverse kinematics.