A three-stage learning algorithm for deep multilayer perceptron (DMLP) with effective weight initialisation based on sparse auto-encoder is proposed in this paper, which aims to overcome difficulties in training deep neural networks with limited training data in high-dimensional feature space. At the first stage, unsupervised learning is adopted using sparse auto-encoder to obtain the initial weights of the feature extraction layers of the DMLP. At the second stage, error back-propagation is used to train the DMLP by fixing the weights obtained at the first stage for its feature extraction layers. At the third stage, all the weights of the DMLP obtained at the second stage are refined by error back-propagation. Network structures an

In recent years, the positioning applications of Internet-of-Things (IoT) based systems have grown increasingly popular, and are found to be useful in tracking the daily activities of children, the elderly and vehicle tracking. It can be argued that the data obtained from GPS based systems may contain error, hence taking these factors into account, the proposed method for this study is based on the application of IoT-based positioning and the replacement of using IoT instead of GPS.  This cannot, however, be a reason for not using the GPS, and in order to enhance the reliability, a parallel combination of the modern system and traditional methods simultaneously can be applied. Although GPS signals can only be accessed in open spaces, GP

Dynamic machine foundations can be considered as a necessary component of the industrial infrastructure. Design of the dynamic equipment foundations has, however, traditionally been grounded on a rule of thumb that is inaccurate and rigid to use at the discretion of the engineers. The conventional rule of thumb, which includes minimum weight ratios and resonance avoidance criteria, has been used singularly with two poles, which can be either conservatively designed systems that are too heavy, or systems that are going to experience too much vibration and fatigue. This paper presents a novel, analytical framework for the reinterpretation of traditional design practices, using a physics-based approach, and results in a single, unified overall

The paper present design of a control structure that enables integration of a Kinematic neural controller for trajectory tracking of a nonholonomic differential two wheeled mobile robot, then  proposes a Kinematic neural controller to direct a National Instrument mobile robot (NI Mobile Robot). The controller is to make the actual velocity of the wheeled mobile robot close the required velocity by guarantees that the trajectory tracking mean squire error converges at minimum tracking error. The proposed tracking control system consists of two layers; The first layer is a multi-layer perceptron neural network system that controls the mobile robot to track the required path , The second layer is an optimization layer ,which is impleme

In this paper, a cognitive system based on a nonlinear neural controller and intelligent algorithm that will guide an autonomous mobile robot during continuous path-tracking and navigate over solid obstacles with avoidance was proposed. The goal of the proposed structure is to plan and track the reference path equation for the autonomous mobile robot in the mining environment to avoid the obstacles and reach to the target position by using intelligent optimization algorithms. Particle Swarm Optimization (PSO) and Artificial Bee Colony (ABC) Algorithms are used to finding the solutions of the mobile robot navigation problems in the mine by searching the optimal paths and finding the reference path equation of the optimal