http://dspace.cityu.edu.hk:80/handle/2031/3771 Tue, 30 Apr 2013 07:48:33 GMT 2013-04-30T07:48:33Z http://dspace.cityu.edu.hk:80/handle/2031/6740 Title: Android camera car – Sengital Ltd. Authors: Lui, Tik On (雷廸安) Abstract: In this project, the main part is, how can we implement video streaming on every android platform devices and how can we make a specified "tablet" devices that we want? In this project, we want to design a device which is designed on embed Android platform based on Linux Kernel to transmit real time video signals to any Android devices. To make it more interesting, we try to embed this system into a simple motor car and it provides the user playing the car with the first-person view likes a driving a car in the real world. For the performance, the "tablet" device runs smoothly and also video streaming is well performed in the good quality of network or high speed CPU environment. To deal with this issue, we have tried modified the program and its performance is finally improved. Apart from that, the motor car can be controlled smoothly by the user without any problems. This project may be useful for the people who want to study about video conferencing and tablet hardware design since we will explain the rationale and procedure behind of these two parts. Notes: Nominated as OAPS (Outstanding Academic Papers by Students) paper by Department in 2012-13. Sun, 01 Jan 2012 00:00:00 GMT http://dspace.cityu.edu.hk:80/handle/2031/6740 2012-01-01T00:00:00Z http://dspace.cityu.edu.hk:80/handle/2031/6735 Title: Symbol synchronization in OFDM Authors: Cheung, Chi Ho (張志豪) Abstract: Orthogonal frequency division multiplexing (OFDM) is a popular transmission technique for high data rate applications. It uses orthogonal subcarriers to transmit data with long symbol duration and thus reduce the effect of inter-symbol-interference due to the multipath channel. However, the successful data detection at the receiver relies on locating the FFT window correctly at the received data sequence ---- the so-called symbol synchronization. In this project, various symbol synchronization techniques are investigated and compared. These include S&C's method, Minn's first method, Minn's second method and Park's method. The performance under an Additive White Gaussian Noise (AWGN), frequency offset, phase offset and multipath channels are investigated. Simulation results and theoretical analysis show that (a) Park's method has the best performance under AWGN channel; (b) all of them are not sensitive to frequency offset and phase offset; (c) they all suffer from the degradation due to multipath channel. In this project, a new symbol synchronization algorithm is proposed. The proposed method uses the combination of Park's method with multiplying the sum of a window on Park's timing metric. The computation complexity is comparable to Park's method. Simulation results show that it has lower estimation mean and variance compared with the above four methods under some multipath channels. Notes: Nominated as OAPS (Outstanding Academic Papers by Students) paper by Department in 2012-13. Sun, 01 Jan 2012 00:00:00 GMT http://dspace.cityu.edu.hk:80/handle/2031/6735 2012-01-01T00:00:00Z http://dspace.cityu.edu.hk:80/handle/2031/6736 Title: 1.8GHz balanced low noise amplifier Authors: Yip, Sui Chun (葉瑞珍) Abstract: A low noise amplifier is important to a receiver in the Wireless Communication. As it is the key component to boost the signal and minimize the noise produced by itself, it can much extend the distance between the transmitter and the receiver. A traditional low noise amplifier achieves low noise but without good input matching. Then the balanced amplifier is a practical method to improve the input matching. This report presents a new balanced low noise amplifier with size reduction. The amplifier was designed at the centre frequency of 1.8GHz, and it consists of three main parts, namely input branch-line coupler, two pieces of well biased transistor, and output branch-line coupler. The input branch-line coupler not only divides signal into two paths with equal magnitude and 90° out-of-phase, but also provides optimum source impedance for the biased transistor to minimize noise figure. Similarly, the output branch-line coupler not only combines signals, but also provides optimum load impedance to maximize gain. Both input and output branch-line couplers were designed with the general port impedance but not purely resistive 50Ω port. It results in non-standard values of both characteristic impedances and electrical lengths of the four branch lines, where its size is similar to the conventional coupler. Conventional balanced low noise amplifier was also designed for comparison. The new design achieves about 36% size reduction compared with conventional design with similar performances. The impedances of both input and output ports are matched, where measured S11=-13.4dB and measured S22=-17.3dB. Measured gain and noise figure are 14.5dB and 2.7dB, respectively. Notes: Nominated as OAPS (Outstanding Academic Papers by Students) paper by Department in 2012-13. Sun, 01 Jan 2012 00:00:00 GMT http://dspace.cityu.edu.hk:80/handle/2031/6736 2012-01-01T00:00:00Z http://dspace.cityu.edu.hk:80/handle/2031/6680 Title: SmartSur: Smart Video SURveillance System Authors: Li, Yiwei (李憶偉) Abstract: In this report, I present the SmartSur system, which is a smart video surveillance system that is capable of detecting a pair-wise event called peope-meet. SmartSur mainly consist of three stages: pre-processing, human detection and event detection. The pre-processing can be decomposed into four major stages: background subtraction, noise elimination and hole filling, contour recognition and object labelling. All these four stages together help to find the moving foreground objects from the raw input video frame. Human detection focuses on finding human from the detected foreground objects which involves 3 major stages, namely HOG feature extraction, SVM training and SVM testing. In the event detection part of the system, a manually defined event called people-meet will be detected based on the position of the detected human. The system is developed on C++ platform using OpenCV as the core library. Besides, the implemented system is tested by the AVSS surveillance video database with a well-recognized result. Notes: Nominated as OAPS (Outstanding Academic Papers by Students) paper by Department in 2012-13. Sun, 01 Jan 2012 00:00:00 GMT http://dspace.cityu.edu.hk:80/handle/2031/6680 2012-01-01T00:00:00Z