Phd thesis artificial neural networks

Phd thesis artificial neural networks

Phd Thesis On Artificial Neural Network When we say that we are offering you reasonable essay service, we are keeping our word of honor which is to give you packages that are light on your pocket. It is entirely up to you which package you choose, whether it is the cheapest one or the most expensive one, our quality of work Phd Thesis On Artificial Neural Network will not depend on the package/10() Artificial Neural Networks: A Financial Tool As Applied in the Australian Market Ph.D. Thesis by Clarence Nyap Watt Tan Bachelor of Science in Electrical Engineering Computers (), University of Southern California, Los Angeles, California, USA Master of Science in Industrial and Systems Engineering () Neural Networks In this section, we will describe neural networks brie y, provide some termi-nology and give some examples. Neural networks are weighted graphs. They consist of an ordered set of layers, where every layer is a set of nodes. The rst layer of the neural network is called the input layer, and the last one is called the output

PhD Projects in Artificial Neural Network (#1 PhD Research Projects)

edu no longer supports Internet Explorer. To browse Academia. edu and the wider internet faster and more securely, please take a few seconds to upgrade your browser. Log In with Facebook Log In with Google Sign Up with Apple.

Remember me on this computer. Enter the email address you signed up with and we'll email you a reset link. Need an account? Click here to sign up. Download Free PDF. A Sensing System for an Autonomous Mobile Robot Phd thesis artificial neural networks on the Paraconsistent Artificial Neural Network Lecture Notes in Computer Science, Jair Minoro Abe.

Download PDF Download Full PDF Package This paper. A short summary of this paper. Download PDF. Download Full PDF Package. Translate PDF. Reasoning-based Intelligent Systems, Vol. t uol. BPS — com João Inácio Da Silva Filho Department of Electric Engineering, Santa Cecilia University — UNISANTA, Oswaldo Cruz street, — Santos City, SP, Brazil Email: inacio unisanta.

br Helga Gonzaga Martins Artificial Intelligence Application Group — GAIA, Federal University of Itajubá, Av. com Abstract: This paper shows the results of a sensing system for an autonomous mobile robot. The sensing system is based on the paraconsistent neural network. The type of artificial neural network used in this work is based on the paraconsistent evidential logic Eτ.

The objective of the sensing system is to inform the other robot components the obstacle position. The reached results have been satisfactory.

Keywords: paraconsistent neural network; paraconsistent logic; autonomous mobile robot; sensing system. Reference to this paper should be made as follows: Torres, C.

and Martins, H. Biographical notes: Cláudio Rodrigo Torres received the PhD and MSc degrees from Itajuba Federal UniversityBrazil, and the BSc and degree from Santa Cecilia University, Brazil, all in Electric Engineering. He works as Professor at Universidade Metodista de São Paulo, phd thesis artificial neural networks, Brazil. Germano Lambert-Torres received the BSc and MSc degrees from Itajuba Federal University, Brazil, and the PhD degree from Ecole Polytechnique at Montreal, all in Electrical Engineering inandrespectively.

He also received a BSc degree in Mathematics from the Itajuba Faculty of Philosophy, Sciences and Letters, both in Currently, he works as full Professor at Itajuba Federal University.

He has more than journals and conference published papers and more than 70 master and PhD supervisions to his credit. His research interests are in the areas of power system operation and intelligent system applications. Phd thesis artificial neural networks is the Coordinator of the Paraconsistent Applied Copyright © Inderscience Enterprises Ltd. Results of a sensing system for an autonomous mobile robot Logic Group GLPA and member of the Group of Logic and Theory of the Science of IEA Institute of Phd thesis artificial neural networks Studies of São Paulo University.

Helga Gonzaga Martins received the BSc and MSc degrees in Mechanical Engineering and the PhD degree in Electrical Engineering from Itajuba Federal University inandrespectively. Her research interests are in the areas of classical logic, paraconsistent logic, and artificial intelligence applications.

In this work, an autonomous mobile robot is considered as a system divided into three other subsystems: planning subsystem, sensing subsystem and mechanical subsystem. The planning subsystem is responsible for generating the sequence of movements the robot must perform to achieve a set point.

The sensing subsystem has the objective of informing the planning subsystem the position of obstacles; and the mechanical subsystem is the robot itself, it means the mobile mechanical platform which carries all devices from the other subsystems. This platform must also perform the sequence of movements borne by the planning subsystem. The planning subsystem and the sensing subsystem have already been implemented, but the mechanical subsystem has not been implemented yet.

The sensing subsystem uses the paraconsistent artificial neural network Da Silva Filho et al, phd thesis artificial neural networks. This type of artificial neural network is based on the paraconsistent evidential logic Eτ. We describe some concepts about Eτ in the next section.

This paper focuses on the sensing system of the robot Torres et al. The μ is a value between 0 and 1 that represents the favourable evidence in which the sentence is true.

The λ is a value between 0 and 1 that represents the contrary evidence in which the sentence is true. Through the favourable and contrary degrees, it is possible to represent the four extreme logic states as shown in Figure 1. Da Silva Filho proposed the paranalyser algorithm, phd thesis artificial neural networks. By this algorithm, it is also possible to represent the nonextreme logic state. Figure 2 shows this. Torres et al. In the next section, we describe the proposed sensing system.

After that, it must inform the other components of the robot the position of the obstacles. The sensing system may get information from any type of sensor. It is shown in this paper that a sensing system is able to deal with information from just one ultrasonic sensor.

But, if there is more than one sensor in the robot, phd thesis artificial neural networks, it is possible to build a sensing system similar to the one shown here for each sensor. A similar method is proposed here, but instead of using probabilistic representation, the paraconsistent evidential logic Eτ is used.

The proposed sensing system aims to generate a favourable evidence degree in each environment position. The favourable evidence degree is related to the sentence: there is obstacle in the analysed position. The sensing system is divided into two parts, phd thesis artificial neural networks. The first part is responsible for receiving the data from the sensors and sending information to the second part of the system.

And the second part is the paraconsistent artificial neural network itself. Figure 3 shows this idea. Figure 3 In the first part of the sensing subsystem, there are also some configuration parameters. They are: 1 Figure 5 shows the distance between the environment coordinates a.

Figure 4 Angle α Figure 5 Distance between coordinates Figure 6 Ultrasonic sensor conical field of view β Representation of the sensing system The proposed sensing system is prepared to receive data from ultrasonic sensors. The robot sensors are on the mechanical subsystem. So, this subsystem must treat the data generated by the sensors and send information to the first part of the sensing subsystem.

The data the mechanical subsystem must send to the first part of the sensing subsystem are: D, α, Xa and Ya. Figure 4 shows the angle α. The first part of the sensing system generates three favourable evidence degrees, μ1, μ2 and μ3. The favourable evidence degree μ1 is related to the distance between the sensor and the obstacle. The nearer the obstacle is to the sensor, the bigger the μ1 value is. Results of a sensing system for an autonomous mobile robot The favourable evidence degree μ2 is related to the coordinate position on the arch BC shown in Figure 6.

The nearer the analysed coordinate is to the point A, the bigger is the μ2 value. And the nearer the analysed coordinate is to the points B and C, the smaller is the μ2 value.

And this probability decreases as we analyse the region near to the points B and C. Eventually, the favourable evidence degree μ3 is the previous value of the coordinate of favourable evidence degree. Below, the cells are described. Phd thesis artificial neural networks there are two configuration parameter inputs Ftct and Ftc, phd thesis artificial neural networks. Figure 7 shows the graphic representation of this cell. It is called phd thesis artificial neural networks certainty interval φthe certainty degree interval that can be modified without changing the uncertainty degree value.

Figure 8 shows the graphic representation of CNAPpa. Figure 8 Graphic representation of the analytic paraconsistent artificial neural cell Ftct 6 2 4. The output 1 S1 is the resultant evidence degree μE. The value of the output S1 is the same as the input μ. But the output value may be limited through the parameter control input Ftc. Figure 9 Figure 10 The graphical representation of the database generated by the first test of the sensing subsystem Phd thesis artificial neural networks chosen paraconsistent neural network architecture for the sensing system 5.

The database stores the favourable evidence degree in each environment position analysed. The result of three tests is shown here. The information from one ultrasonic sensor was considered as the sensing system inputs.

The distance between the environment coordinates a is The angle of the ultrasonic sensor conical field of view β is The number of positions on the phd thesis artificial neural networks of the sensor conical field of view considered by the system n is

PhD Projects in Neural Networks - PhD Thesis in Neural Networks

, time: 0:35

Phd Thesis Artificial Neural Networks thing about essay writing is that requires more than just the ability to write well (which could be a struggle on its own for some students). Proper paper writing includes Phd Thesis Artificial Neural Networks a lot of research and an ability to form strong arguments to defend your point of view. It also requires knowledge about how to present your /10() The sensing subsystem uses the paraconsistent artificial neural network (Da Silva Filho et al., , ). This type of artificial neural network is based on the paraconsistent evidential logic (Eτ). We describe some concepts about Eτ in the next section Thesis On Artificial Neural Networks with cheap prices is to make sure that you get a quality paper with original and non-plagiarized content. 50% off on all orders. ORDER NOW Toggle navigation. Limited Time Only! Please note. Orders of are accepted for higher levels only Thesis On Artificial Neural Networks (University, Master's, PHD)/10()