Journal of Majlesi Journal of Mechatronic Systems
Volume:4 Issue: 4, Dec 2015

Ability to predict direction of stock/index price accurately is crucial for market dealers or investors to maximize their profits. Data mining techniques have been successfully shown to generate high forecasting accuracy of stock price movement. Nowadays, instead of a single method, traders need to use various appropriate techniques to gain more information about the future of the markets. In this paper, three different techniques of data mining are discussed and applied to predict price of Tehran stock market. The approaches include decision trees (CART and CHAID) and Neural Network that execute on Saipa, Iran Khodro, Telecommunication, Mapna and Saderat Bank datasets. The aim of this paper is generating an effective predicting model to forecasting future price in Tehran stock market. Price prediction in stock market helps investors for exact and quick identification and more investment on valuable portion. Finally it causes portion basket of investors optimized. Results show that the Neural Network is suitable for producing industries and CART is suitable for services industries. Next, many rule induction that achieved CART modeling, discussed.

In this paper, the Parallel Distributed Compensation (PDC) method is used to stabilize a cart-mounted inverted pendulum and Overhead Crane model. One of the significant issues of using PDC approach for systems with nonlinear terms is finding a linear sectors. Since equations of motion (EoM) of an inverted pendulum and overhead crane include some complicated nonlinear terms, finding sectors is often impossible. Therefore, the traditional PDC has some difficulties from the practical point of view. In order to overcome such a problem, here, two different approaches have been proposed. In the first approach, prior to utilizing a PDC, complex equations have been simplified using feedback linearization method. PDC method is then applied to the obtained closed loop system. In the second strategy, named as the approximated PDC, after eliminating the small terms in the EoM, PDC method is applied. The results of simulations pertinent to PDC controller and traditional LQR method have been compared. Finally, for verification of the presented approximated approach, practical implementation on the experimental crane setup has been done and results reported and compared with simulation.

How to reconfigure a logic gate is an attractive subject for different applications.Chaotic systems can yield a wide variety of patterns. Here, we use this feature for produce a logic gate.This feature forms the basis for designing a dynamical computing device that can be rapidly reconfigure to become any desired logic operator.This logic gate that can reconfigure to any logic operator when placed in its chaotic state, called chaotic logic gate.The reconfiguration realize by set the parameter values of chaotic logic gate.In this paper we present mechanisms about how to produce a logic gate based on the logistic map.Furthermore,we propose a reconfigurable structure based on a chaotic system and genetic algorithm is used to set the parameter values.

In this paper, an appropriate method is presented to data transfer speed and slaves distance from master. Speed controlling and torque monitoring of servo motors (5pcs) were done by Labview software, and also master and slave connection was performed by Modbuse protocol. By writing a program based on labview for data exchange in modbus protocol, archive automation and store sampling values, torque and speed of servo motors were monitored. In this program, communicating flowchart with servo motor drives and sending of commands in modbus protocol and receiving of torque values (analogue values) were created by PCI_GPAD1916 card. Also, programming in Labview software were done based on the created flowchart. The results showed that this method has high speed sampling and signal processing.

In this paper, synchronization problem of the chaotic dynamical systems in the presence of unknown input is studies. Hence is presented, a sliding mode controller design method for synchronization of two identical hyper chaotic system with an unknown inputs and different initial conditions. Using the theory of Lyapunov, stability control system is proven. The simulation results show that the purposed controller is able to synchronization error with unknown inputs are fully converge to zero and is suitable for application to a larger class of chaotic. The control signal is low that changes are the ability to build and implement.

This paper presents a novel control approach for a lower limb exoskeleton based on adaptive neuro-fuzzy inference system (ANFIS). An exoskeleton is a wearable robot, usually with an anthropomorphic configuration, capable of tracking the movements of the user’s extremities. The goal is to design a controller such that robot can track the human’s motion without the human feeling an interaction force. A ten Degrees of Freedom lower limbs exoskeleton is considered to study the performance of the proposed controller. The system is actuated by four pneumatic actuators which are located in hip and knee joints. For each actuator, a separate ANFIS is trained, witch control exoskeleton in joint space based on the angles of the knee and hip joints in sagittal plane. The simulation results demonstrate that proposed controller is tracked the human’s motion with minimal error.

For many engineering problems, a buoyant jet is commonly used. It possesses both momentum and buoyancy, and it is usually discharged at an inclined angle, so asymptotic solutions for vertical jets and plumes are inadequate. To obtain the mixing characteristics of an inclined buoyant jet (such as the jet path and jet concentration), a general jet integral model is often used. In this paper a triple multilayer perceptron is used for analyzing the length and height trajectory of a typical buoyant jet boundary. Experimental results shows that using triple MLP for each output is better that single MLP.