Design and simulation of intelligent controller for robot arm

response enhancement.

A PROJECT PROPOSAL

Submitted by

ABUBAKARSADIQ ALI MUHAMMAD

SPS/15/MEE/00022

[Link] CONTROL AND INSTRUMENTATION

SUPERVISOR: DR HASSAN BASHIR

Department of Electrical Engineering

Faculty of technology

Bayero University, Kano.

The wider application of automatic control has developed rapidly in recent years.
The reason for this is the complexity of modern plant and the constraints imposed
by the increasing demand for higher quality products. Hence, the design of a
controller which possesses learning capability becomes highly desirable. Robots
are highly reliable, dependable and technologically advanced factory equipment.
For robots to become effectively used in a wide range of applications, they must
gain the ability to work in unpredictable and changing environments. Robots used
in space exploration and construction will have to sense their environment and
carry out their tasks regardless of the presence of objects in the work area. The
ability to move an arm around an obstacle and to a goal is an intuitive skill for
human beings. Translating that skill into instructions for a robotic arm, however,
is not an easy task [1].

The ability of a machine to emulate human behavior has always been the goal of
artificial intelligence. Neural networks and fuzzy logic systems are two of the most
important results of research in the area of artificial intelligence [3]. They have
been effectively applied to everything from voice and image recognition to
toasters and automobile transmissions. Neural networks are best known for their
learning capabilities. Fuzzy logic is a method of using human skills and thinking
processes in a machine. More than 90% of industrial controllers are still
implemented based around proportional-integral-derivative (PID) algorithms,
particularly at lowest levels. With its three-term functionality covering treatment
to both transient and steady-state responses, proportional-integral-derivative
(PID) control offers the simplest and yet most efficient solution to many real-
world control problems. The PID controller is a one of the earliest industrial
controllers. It has many advantages: Its cost is economic, simple easy to be tuned
and robust. This controller has been proven to be remarkably effective in
regulating a wide range of processes [4].
1.2 Problem Statement

In recent years, robots are very widely used in practical applications. The
weakness point of previous methods is determining acceptable trajectories, and
its need long time to obtain a solution. This intelligent approach will be used to
enhance the output response of the system, since almost research in recent years
integrates the intelligent controllers to obtain modified controller.

Conventional controller has issues in handling the time response, overshoot and
steady-state error of a robot manipulator, this issues can be mitigate using
nonlinear controller.

Aim and objectives

The aim of this project is to design and simulate an intelligent controller for the
enhancement of robot arm response. The objectives are

1 To design a single link robot arm model.

2 To design a separated fuzzy and neural controller.

3 To design a PID controller as a selector switch to track the system on line.

4 To investigate the designed controller via simulation.

Methodology

The methods to be adopted in this research are represented in the following

The mathematical model of the single link robot arm will be obtained and the
result will be represented using state space analysis.

Two separate controllers’ i.e fuzzy logic and neural network controller will be
designed using MATLAB/SIMULINK.

PID controller will be designed and the PID parameters will be obtained using
Ziegler Nichols closed loop method.

Simulation of the design will be done using MATLAB/SIMULINK and comparison of

Expectation

The robotic arm designed using intelligent controller is expected to have a better
overshot, settling time, and steady state error compare to conventional
controllers.

Literature Review

Robotic arm just like the human arm consist of rigid links interconnected by
flexible joint and the analytical study of the geometry of motion of mechanism is
known as Kinematics. Kinematics is the first step towards robotic control. The
accurate control of motion is a fundamental concern in engineering applications,
where placing an object in the exact location with the exact possible amount of
force and torque at the correct exact time is essential for efficient system
operation.

Experimental data suggests that the desired trajectory of the arm end point is
planned in the task-oriented (visual) coordinates and is independent to the joint
coordinates.

The control strategies for robotics are classified into classical and non-classical
methods. Control robot arm using classical methods is conventional and have a
lot of problems because robotic systems are always highly nonlinear unlike
artificial control (fuzzy logic, genetic algorithm, neural network). Fuzzy logic have
shown to be able to approximately model any complex and nonlinear system,
capable of modeling vagueness, handling uncertainty and supporting human type
reasoning. But, it has the drawback of learning and depends on expert
knowledge. Neural network have been made to automate knowledge acquisition
for fuzzy systems by means of learning. The performance of any fuzzy system
mainly depends on the input-output memberships functions, the if-then rules,
and tuning of both. The choice of defuzzification method is another factor, which
also influences the performance. Unfortunately, there are no formal methods to
define the membership functions. Genetic Algorithm is the suitable choice for
optimization where no a priori information about the membership function and
rule-base is available.

This research has an ultimate goal of Designing, Modelling and Control of a robot
arm using a separated neural, fuzzy and PID that can be used in the production
industries.

Conclusion

The intelligent control technique is very useful technology in robot system, since
they have the ability to learn how to control a process from sample data. In this
paper, separated Neuro-PID-Fuzzy controller is proposed to emulate a human’s
example control of robotic arm.

References

[1] Kellyn W. E. (1994). "Neuro-Fuzzy Control of a Robotic Arm " MSc. Thesis,

electrical engineering , Texas Unv. August.

[2] Cirstea M.N, Dinu A., Khor J.G., and Mcormick M. (2002). " Neural and Fuzzy
Logic Control of Drives and Power Systems" , Vol. No. 22, pp. 13-15.

[3] D. Bin Dulaini.(2014) “modeling and control of 6DOF industrial robot using
fuzzy logic controller” [Link] thesis, Universiti Tun Hussein Onn Malaysia.

[4] Ahmed Karim Abdallah Albakri “Design and Simulation Intelligent Controller
for Robot Arm Response Enhancement Using Intelligent Techniques” Technical
Collage / Almusyeib.
[5] Kazuo Kiguchi “Neuro-Fuzzy Control of a Robotic Exoskeleton

With EMG Signals” IEEE TRANSACTIONS ON FUZZY SYSTEMS, VOL. 12, NO. 4,
AUGUST 2004.

[6] Farhan A. Salem, Modeling, Simulation and Control Issues for a Robot ARM;
Education and Research (III). I.J. Intelligent Systems and Applications, 2014, 04,
26-39 Published Online March 2014 in MECS.