DSpace collection: 專書之單篇

結合「競賽導向學習」與「計畫導向學習」之淡江大學電機系團隊的智慧機器人研發

Mon, 13 Apr 2020 04:10:35 GMT

title: 結合「競賽導向學習」與「計畫導向學習」之淡江大學電機系團隊的智慧機器人研發 abstract: 淡江大學電機系的智慧機器人研發團隊目前是由電機系翁慶昌特聘教授、李世安副教授以及劉智誠助理教授等三位老師帶領一些博碩士研究生以及大學部專題生所組成的團隊。本文將以團隊的背景、團隊在智慧機器人的研發歷程與成果、團隊近期的研究與發展以及結語等4個部分來介紹團隊是如何結合「競賽導向學習」與「計畫導向學習」在智慧機器人的研發上。

A GPU-Accelerated Adaptive Simultaneous Dynamic Range Compression and Local Contrast Enhancement Algorithm for Real-Time Color Image Enhancement

Thu, 02 Mar 2017 18:10:52 GMT

title: A GPU-Accelerated Adaptive Simultaneous Dynamic Range Compression and Local Contrast Enhancement Algorithm for Real-Time Color Image Enhancement abstract: Dynamic range compression is an important function used in modern digital video cameras and displays to improve visual quality of standard dynamic range color images. This chapter presents a real-time implementation of an adaptive contrast-enhancing image dynamic range compression algorithm on a graphics processing unit (GPU) for color image enhancement. To achieve this purpose, an image-dependent nonlinear intensity transfer function is first presented to produce a satisfactory dynamic-range compression result with less color artifacts. The proposed algorithm is then derived by combining the proposed nonlinear intensity transfer function with an existing simultaneous dynamic range compression and local-contrast enhancement (SDRCLCE) algorithm, which is a parallelizable method to compress image dynamic range while enhancing local contrast of output images. Finally, the proposed algorithm is implemented on the GPU by using NVIDIA Compute Unified Device Architecture (CUDA), achieving real-time performance in processing high-resolution color images. The proposed GPU-accelerated color image enhancement method had been implemented on a NVIDIA NVS 5200M GPU. Experimental results show that the proposed GPU implementation gains about 7 times acceleration, including the cost of memory copy between host and device, compared with a LUT-accelerated implementation on an Intel Core i7-3520M CPU for color images with size 1024 × 1024 pixels.

Image Reconstruction for Buried Inhomogeneous Dielectric Cylinders by Evolutionary Algorithms

Fri, 15 Mar 2013 08:25:54 GMT

title: Image Reconstruction for Buried Inhomogeneous Dielectric Cylinders by Evolutionary Algorithms

Electromagnetic imaging of Buried Perfectly Conducting Cylinders TArgets Using APSO

Fri, 15 Mar 2013 08:23:53 GMT

title: Electromagnetic imaging of Buried Perfectly Conducting Cylinders TArgets Using APSO

Pervasive Health-Care Device Industry Analysis: Focusing on Sleep-Treatment Applications

Tue, 12 Mar 2013 06:10:11 GMT

title: Pervasive Health-Care Device Industry Analysis: Focusing on Sleep-Treatment Applications abstract: In this chapter, we want to determine the factors and risks that affect the viable business model for pervasive healthcare systems - with an emphasis on sleep treatment devices. We used a hybrid analysis utilizing, i) Porter’s Five Forces, to determine the competitive advantages of the pervasive healthcare industry and ii) Delta Model, to determine the three strategic options related to customer bonding and adaptive processes which support these options. We have found that with the hybrid analysis pervasive healthcare industry will be highly competitive and strive to meet customer needs making it the next big trend in the medical industry overall. In particular, the ability to customize accordingly to the patients individual conditions is a key differentiator in such an industry. In addition, the major players and value chain of the sleep treatment industry are illustrated.

Multi-Robot Cooperative Sensing and Localization

Mon, 11 Mar 2013 06:21:07 GMT

title: Multi-Robot Cooperative Sensing and Localization abstract: This paper presents a novel method for mobile robot localization based on multi-robot cooperative sensing. A multi-robot cooperative localization scheme is proposed using visual detection and sensor data fusion. The visual detection system employs a stereo vision module for both observing other robots and obtaining environmental information. Each mobile robot is able to recognize its teammates by using the vision system. The localization error is reduced through the proposed sensor fusion algorithm. The cooperative localization algorithm consists of two stages: serial fusion and parallel fusion. Serial fusion aims to identify the positional uncertainty of an observed robot while parallel fusion reduces its positional error based on Kalman filtering. The multi-robot cooperative localization system has been realized through the server-client architecture. Experimental results are presented to validate the effectiveness of the proposed algorithms.

Hybrid Path Planning Incorporating Global and Local Search for Mobile Robot

Fri, 25 Jan 2013 02:40:13 GMT

title: Hybrid Path Planning Incorporating Global and Local Search for Mobile Robot abstract: This paper proposes a hybrid path planning algorithm incorporating a global and local search mechanism for mobile robot. The global path planning is based on Voronoi graph to establish a backbone path for the map with significantly reduced nodes for the original grid map. With the use of the backbone path, the D* algorithm is adopted to determine a shortest path between the starting and end points. Taking advantages of the D* algorithm and Voronoi graph, the proposed hybrid path planning algorithm is capable of obtaining a desired path for the mobile robot, overcoming the efficiency problem while maintaining maximum safety distance from the obstacles when the mobile robot navigates in the environment.

Hardware accelerator design for image processing

Fri, 25 Jan 2013 02:28:31 GMT

title: Hardware accelerator design for image processing abstract: This paper proposed an image processing system based on hardware accelerator design method in FPGA chip. The architectures of the image system will be described. A hardware accelerator for scaling image is designed with Avalon-MM burst mode in System-on-a-Programmable-Chip (SOPC). There is a human-machine interface which display on the LCD Touch Panel Module (LTM) is designed in the image processing system. A user can choose a scaling factor by touching screen panel, and then LTM will display a 640 × 480 image on the LTM screen. Finally, the experimental result shows the comparison of the two design methods, and the hardware accelerator has better performance than Nios II processor.

Design and Implementation of Omni-directional Walking System for Humanoid Robot

Mon, 21 Jan 2013 08:29:14 GMT

title: Design and Implementation of Omni-directional Walking System for Humanoid Robot abstract: An omni-direction walking system for humanoid robot is presented in this paper. This method combines a mark time motion and a walking trajectory planning system. Three sub-motions, including leg shifting, leg lifting and leg opening, with same frequency are proposed to accomplish the mark time motion. The proposed omni-directional walking system is established to make robot walk fast in the soccer field.

Motion and Emotional Behavior Design for Pet Robot Dog

Mon, 21 Jan 2013 08:23:03 GMT

title: Motion and Emotional Behavior Design for Pet Robot Dog abstract: A pet robot dog with two ears, one mouth, one facial expression plane, and one vision system is designed and implemented so that it can do some emotional behaviors. Three processors (Inter® Pentium® M 1.0 GHz, an 8-bit processer 8051, and embedded soft-core processer NIOS) are used to control the robot. One camera, one power detector, four touch sensors, and one temperature detector are used to obtain the information of the environment. The designed robot with 20 DOF (degrees of freedom) is able to accomplish the walking motion. A behavior system is built on the implemented pet robot so that it is able to choose a suitable behavior for different environmental situation. From the practical test, we can see that the implemented pet robot dog can do some emotional interaction with the human.

Motion and Emotional Behavior Design for Pet Robot Dog

Thu, 29 Mar 2012 08:37:39 GMT

Threaded C and Freezer OS

Thu, 29 Mar 2012 08:36:41 GMT

title: Threaded C and Freezer OS abstract: Threaded C is a meta-language that is based on C, but is annotated with thread, monitor thread, and semaphore markup. Threaded C uses the runtime provided by the Freezer OS, a small, memory-efficient embedded kernel. The combination of Freezer OS and Threaded C allows the simple expression of common control problems in robotics. The system is geared especially towards robotics education, as it matches the mental map that children have of how control structures should work.

Learning of Facial Gestures Using SVMs

Thu, 29 Mar 2012 08:35:36 GMT

title: Learning of Facial Gestures Using SVMs abstract: This paper describes the implementation of a fast and accurate gesture recognition system. Image sequences are used to train a standard SVM to recognize Yes, No, and Neutral gestures from different users. We show that our system is able to detect facial gestures with more than 80% accuracy from even small input images.

Fuzzy Chaotic Synchronization and Communication– Signal Masking and Encryption

Mon, 15 Aug 2011 05:40:48 GMT

title: Fuzzy Chaotic Synchronization and Communication– Signal Masking and Encryption abstract: In this chapter, we address synthesis approaches for signal synchronization and secure communications of chaotic systems by using LMIbased fuzzy system design methods. Following a general form of TakagiSugeno fuzzy chaotic models, the structure of the response system is firstly proposed. Synthesizing from the observer and controller points of view, two developed drive-response systems achieve asymptotic synchronization. For chaotic communications, the asymptotical recovering of messages is ensured by the same framework. As a further application, a chaotic cryptosystem is proposed. The synchronization design is applied to transmit and decrypt the ciphertext. The combination of both chaotic signal and cyptosystem characteristics achieves communication with a higher level of security. Finally, several well-known chaotic systems are used in numerical simulations and DSP-based experiments.

Security Robot Simulator

Wed, 10 Aug 2011 03:44:56 GMT

title: Security Robot Simulator abstract: Building intelligent behaviors is an important aspect of developing a robot for use in security monitoring services. Simulating and testing the robot behavior in a virtual environment prior to producing the robot and then conducting practical experiments can greatly reduce the cost and duration of the testing process. This research proposes a framework for the simulation of security robots, called the security robot simulator (SRS), which is aimed at providing a fully inclusive simulation environment from fundamental physics behaviors to high-level robot scenarios for developers. Human simulation is also integrated into the robot simulator for simulating interactions between the security robot and human personnel. The simulator was implemented in Microsoft Robotics Developer Studio (MSRDS), a services oriented robotics platform composed of a simulation core and four decentralized modules: scenario event, patrol planner, robot unit, and civilian modules. The results show that the four modules fulfill the requirements of a security robot.

提高光網路穩定度之光性能監視技術

Fri, 26 Mar 2010 14:20:16 GMT

title: 提高光網路穩定度之光性能監視技術

提高光網路穩定度之光性能監視技術

Fri, 26 Mar 2010 14:20:08 GMT

title: 提高光網路穩定度之光性能監視技術

Image reconstruction of buried biaxial dielectric cylinders

Fri, 26 Mar 2010 14:19:55 GMT

title: Image reconstruction of buried biaxial dielectric cylinders

Identification of singleton fuzzy models via fuzzy hyper-rectangular composite NN

Fri, 26 Mar 2010 14:18:30 GMT

title: Identification of singleton fuzzy models via fuzzy hyper-rectangular composite NN