Animatronic Eye Tracking Mechanism for Real Human Eye Motion for Humanoid Robots
Assoc. Prof. Dr. Eng. Amir R. Ali
Eng. Malek Mahmoud
Ahmed Gamal, B.Sc. (2020)
Brief description for the project:
This research revolves around the mechanical, electrical, and software design for eye-mechanism for a robot. As tools for dynamic system modeling both conventional methods such as transfer function or state-space representation and modern power flow-based methods are available. Robots are at the situation to turn into our ordinary allies sooner rather than later. All things considered numerous obstacles should be cleared to accomplish this objective. One of them is the reality that robots are as yet not ready to act like a real human with at least two of the basic concepts first for moving body normally and the second with our topic in this paper for eye-mechanism. For building an instrument that recognizes the genuine human movement itself with eye system following for a robot, it will help robots a great deal to be progressively practical acting like a genuine human when somebody is conversing with him or simply wave. For humans, it will be all the more fulfilling for the inclination that the robot responding with him. This bachelor thesis develops control and a mathematical model of the eye-mechanism with the different method to analysis the sketch to the specific method which called the Port-Hamiltonian system. For a little assistance with modeling and simulation to enter models graphically, like drawing a designing plan and to analysis equations with graphs results utilizing Port-Hamiltonian system. The upside of the Port-Hamiltonian is that they can coordinate various kinds of the dynamic system, the controller and the motor can demonstrate modeled, simulated and reproduced out and out in a similar procedure. The Port-Hamiltonian will be set up for each inflexible body with body-fixed arrange reference outlines, which are associated with parasitic components (damping and consistence) to one another. Port-Hamiltonian are demonstrated to be successful in comprehending dynamic as well as kinematic issues.
Finally, the model is used to simulate the dynamic response using a simulation software package. The simulated results are compared to experimental results and found to have good correlation. The model is suitable for use with a simulation-based design structure.
Challenges of the project:
Enhancing the capability of the models is the most direct way to enhance the capability of the supervisory control system. Therefore, future work on this approach should be concentrated on resolving the problems of the models. Plan to further explore the relationship between our self-supervised continuous grasping approach and reinforcement learning, in order to allow the methods to learn a wider variety of grasp strategies from large datasets of robotic experience.