Showing papers by "Zeng-Guang Hou published in 2005"

Support vector machines (SVM) for color image segmentation with applications to mobile robot localization problems

Abstract: In autonomous mobile robot industry, the landmark-based localization method is widely used in which the landmark recognition plays an important role. The landmark recognition using visual sensors relies heavily on the quality of the image segmentation. In this paper, we use seat numbers as the landmarks, and it is of great importance to the seat number recognition that correctly segment the number regions from images. To perform this assignment, the support vector machine method is adopted to solve the color image segmentation problems because of its good generalization ability. The proposed method has been used for the mobile robot localization problems, and experimental results show that the proposed method can bring robust performance in practice.

An agent-based holonic architecture for reconfigurable manufacturing systems

Abstract: Holonic architectures are more suitable for reconfigurable manufacturing systems compared with hierarchical and heterarchical architectures. A holonic architecture is proposed for reconfigurable manufacturing systems based on the well-known reference architecture PROSA. Considering the special status of the configuration in reconfigurable manufacturing systems, configuration holon is introduced besides the basic holons in PROSA. The basic structure of this holonic architecture, the details of basic holons and cooperation of holons are described in detail. Finally an agent-based holon model is introduced for the realization of the proposed holonic architecture.

A neural network-based camera calibration method for mobile robot localization problems

Abstract: To navigate reliably in indoor environments, a mobile robot has to know where it is The methods for pose (position and orientation) estimation can be roughly divided into two classes: methods for keeping track of the robot's pose and methods for global pose estimation [1] In this paper, a neural network-based camera calibration method is presented for the global localization of mobile robots with monocular vision In order to localize and navigate the robot using vision information, the camera has to be first calibrated We calibrate the camera using the neural network based method, which can simplify the tedious calibration process and does not require specialized knowledge of the 3D geometry and computer vision The monocular vision is used to initialize and recalibrate the robot's pose, and the extended Kalman filter is adopted to keep track of the mobile robot's pose

Identification of motifs with insertions and deletions in protein sequences using self-organizing neural networks

Abstract: The problem of motif identification in protein sequences has been studied for many years in the literature. Current popular algorithms of motif identification in protein sequences face two difficulties, high computational cost and the possibility of insertions and deletions. In this paper, we provide a new strategy that solve the problem more efficiently. We develop a self-organizing neural network structure with multiple levels of subnetworks to make an intelligent classification of the subsequences obtained from protein sequences. We maintain a low computational complexity through the use of this multi-level structure so that the classification of each subsequence is performed with respect to a small subspace of the whole input space. The new definition of pairwise distance between motif patterns provided in this paper can deal with up to two insertions/deletions allowed in a motif, while other existing algorithm can only deal with one insertion or deletion. We also maintain a high reliability using our self-organizing neural network since it will grow as needed to make sure all input patterns are considered and are given the same amount of attention. Simulation results show that our algorithm significantly outperforms existing algorithms in both accuracy and reliability aspects.

Neural network methods for the localization and navigation of mobile robots

Abstract: Neural networks, including support vector machines (SVMs), and principal component analysis (PCA) etc., and their applications have received increasing attentions from many fields such as information processing and systems control. In this paper, we study some useful features of the neural networks, support vector machines and principal component analysis, and apply these methods to the information processing, localization and navigation problems of mobile robots on the basis of the information acquired with multiple sensors, such as visual, ultrasonic and infrared sensors. A Hopfield-type neural network scheme is proposed for real-time optimization and navigation applications. To recognize the doorplate numbers and human faces, support vector machine is proposed for the vision system of the mobile robot. The principal component analysis network is used to process the data acquired by the ultrasonic and infrared sensors to obtain the location and orientation information of the robot. The principal component network can also give feasible directions for movements