Vision-Based Control of a Quad-rotor Manipulation System

Abstract

Using teleoperation process to drive the quadrotor Manipulation System to pick and place the object in a critical application is very difficult and slow process because of human capabilities so that one of the possible solutions to overcome this problem is to use vision-based control to automatically pick and place the object. In this paper we presented the vision-based control of a quadrotor manipulation system (QMS) that consists of 2-serial link robotic arm attached to the bottom of a quadrotor and a camera attached to end-effector which acts as a vision sensor to form eye-in-hand architecture. Image-Based Visual Servoing (IBVS) is used to produce a desired task space trajectory to the inverse kinematics model that generates a desired quadrotor/joint space trajectory for PD-controller to automatically pick and place the object. The proposed system uses 6 actuators only to perform the required 6-DOF trajectory. This minimum number of actuators is useful to facilitate carrying high payload. The vision-based control of a quadrotor manipulation system of the proposed system was designed and tested using MATLAB/SIMULINK program. The simulation results indicate that this system is able to perform automatic pick and place task for 3D objects.