The investigation of circular path generation and hand motion tracking problem using 5DOF vision based robot

Rasid, Farah Amirah (2017) The investigation of circular path generation and hand motion tracking problem using 5DOF vision based robot. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Preview

Text (Full Text) The investigation of circular path generation and hand motion tracking problem using 5DOF vision based robot.pdf - Submitted Version Download (3MB) | Preview

Abstract

Robotic rehabilitation has widely being used especially for upper limb impairment due to neurological disorder or accident. Research has shown that robotic training can also provide repetitive movement without therapist assistant. This project address two main problem in robotic rehabilitation applications which are generating a circular path in Cartesian space and the problem in identifying the positions (x,y,z) of robot’s arm end effector when generating a circular motion that can vary depending on the vision based feedback of patient hand. The aim of this project is to analyse the problem of generating circular path and tracking method using a vision based robot. Secondly to develop circular path algorithm and vision based circular tracking algorithm to match with the patient hand motion and to validate the smoothness of circular path and the accuracy of a vision based robot motion when tracking and following the patient hand motion using statistical method. The idea to generate circular path is by using point to point method in Cartesian space. To get the circular motion, the trigonometry formula combined with formula of circumferences was used to generate the circle. MATLAB is used for the circular generation while the simulation for circular path is done in VREP by using LUA. For the visual part, blob detection, centroid detection (x,y) is used with the imposing of inverse kinematic to track the image. The OpenCV library was used to do blob detection. The robot simulation is done using VREP and the algorithm is developed using LUA. In generating a circular path, the smoothness of circle is important. To measure the smoothness of circle path, the number of point to point is increased. For the tracking part, the accuracy of robot is measured using the statistical method. The youBot have high accuracy in tracking part which is 97% with absolute error 0.03. In conclusion, a vision based robot is reliable for circular hand tracking problem for rehabilitation purposes. A future research can be done to make a precise robot motion by developing a new control method for the robot and add features that enable the robot to save the results of motion for therapist to follow up the patient hand condition.

Actions (login required)

View Item