Thesis Abstract and Dissertation Abstracts
PLANITIA - A Real-Time Fuzzy-Logic Tilt Controller
Abstract Category: I.T.
Course / Degree: B.Sc. (Hons) I.T. (Computer Science)
Institution / University: University of Malta, Malta
Published in: 2003
Background:
Cybernetics is referred to as “the art of steermanship” and it is an inter-disciplinary science that gave rise to the modern field of robotics. One of the major fields of cybernetics involves the integration of Engineering with Biology. This field deals with automatic control devices in machines that basically refer to systems (usually electro-mechanical) that can be given a target goal-state by a human operator, and without any further human intervention, they are able to reach the target state by modifying their own behavior. One of the first cybernetic machine implementations was the centrifugal governor that James Watt had fitted to his steam engine, which can be regarded as a major contributor that led to the industrial revolution. These systems usually make use of Feedback Loops to analyze their output in order to adapt their behaviour for reaching a goal.
The integration of AI techniques into cybernetics enabled a new generation of Robotics to be born with the goal of creating intelligent systems that can offer more complex functionalities. Robotics can be regarded as an intelligent connection between perception and responsive action. This subject is more of an art than a science since a roboticist must be a generalist who is able to incorporate Mechanical Engineering, Electrical Engineering, Computer Science and Artificial Intelligence techniques to produce an autonomous (self-governed) system – a robot. The aim of this project was to solve a practical problem using the techniques in Robotics as well as techniques in the field of Computer Science.
Objectives:
The problem that was analyzed in this project deals with neutralizing the horizontal forces of acceleration induced by an object. The aim was to design a Real-Time system which is able to sense the forces of acceleration on an object standing on a platform and use these acceleration readings to adjust the angle of tilt of this platform in real-time. Effectively the result would be that the object on the platform does not experience any horizontal acceleration even when the platform is in motion. In other words the system should allow say a glass of water to remain on the platform without tipping even though the platform base is moved around or tilted to a side.
An example application for such an automated device could be for holding drinks (without tipping) inside a traveling car. The device would sense any acceleration / deceleration of the vehicle and compensate for this acceleration by tilting the platform appropriately as shown in the diagram. It could also be used for designing tables in marine crafts that would bear able to smoothen out the tilt of the boat. Such an automated system would allow objects on the table to remain stationary even in rough weather when the boat is swinging from side to side.
Methodology:
We know that every object lying on the Earth's surface experiences a constant force of acceleration due to the Earth's gravitational pull. This gravitational constant known to be 9.8 m/sec2 (equivalent to 1g) is a force that ‘pulls’ objects directly towards the centre of the earth. This means that acceleration due to gravity is a force acting in perpendicular to the earth's surface. By using this force of gravity engineers were able to design sensors that measure the angle of tilt of an object with respect to the earth’s surface. A type of sensor that is able to accurately measure tilt (as well as acceleration) is the digital accelerometer.
The project utilizes a dual-axis digital accelerometer board, the ADXL202EB, which is mounted on the top platform. This platform is connected to the base through a shaft & ball-joint which enables it to tilt a maximum of around ±50 degrees on each axis.
Two DC motors (one for each axis) are fixed to the base and through a system of pulley & strings they can change the tilt of the top platform to any angle in either of the X/Y plane. Each of the motor’s polarity/direction is controlled by an LMD18200 H-Bridge driver IC.
Both the motor circuitry and the accelerometer board are interfaced to a PC through the LPT and COM ports respectively. A Real-Time system running on the PC gets tilt/accelerations readings from the accelerometer, computes the current angle of tilt and outputs a PWM (pulse width modulation) signal to the motor circuitry in order to control the speed and direction of each motor to set the top platform to a target angle.
The process is accomplished through a high-frequency Closed Feedback Control Loop that uses an FLC (Fuzzy Logic Controller) algorithm to calculate the appropriate output signals for each motor in real-time. The software makes use of a multi-threaded technique for separating the components of the controller engine. The 3 main components include:
• an input sampling thread that continually checks the status of the hardware sensors
• a feedback control thread that uses the input data to calculate the appropriate output signals
• an output signaling thread that keeps a continuous flow of PWM signals to the motors
Results and Achievements:
Although the prototype and accompanying real-time application software was not designed for industrial use, it enabled a detailed study of the issues involved in integrating AI techniques with an engineering solution for such a practical problem.
The prototype was tested using both a conventional crisp-logic design as well as a fuzzy-logic engine. This enabled an analysis of the performance obtained by both systems. Although the system produced satisfactory results when tested with a crisp-logic controller, the aim of introducing a fuzzy controlled engine was mainly for tweaking performance and obtaining smoother results that would enable such a system to be used for more practical real-world applications.
Thesis Keywords/Search Tags:
Fuzzy Logic Real Time Computer Science
This Thesis Abstract may be cited as follows:
Fenech, K. (2003), PLANITIA - A Real-Time Fuzzy-Logic Tilt Controller, Unpublished B.Sc. (Hons) I.T. (Computer Science) Dissertation, University of Malta
| Thesis Images: | |
|
An ACAD design of the hardware prototype (click to enlarge) |
|
Illustrating the effects of Tilt & Acceleration on an object standing on the platform. (click to enlarge) |
Submission Details: Thesis Abstract submitted by Keith Fenech from Malta on 26-Sep-2003 23:24.
Abstract has been viewed 3298 times (since 7 Mar 2010).
Keith Fenech Contact Details: Email: Keith@RainbowMalta.com
Disclaimer
Great care has been taken to ensure that this information is correct, however ThesisAbstracts.com cannot accept responsibility for the contents of this Thesis abstract titled "PLANITIA - A Real-Time Fuzzy-Logic Tilt Controller". This abstract has been submitted by Keith Fenech on 26-Sep-2003 23:24. You may report a problem using the contact form.
© Copyright 2003 - 2024 of ThesisAbstracts.com and respective owners.