Browse
Recent Submissions
Item Neural network based optimal placement of base stations(UMT.Lahore, 2010) Waqas Tariq ToorNeural network has been used as an optimization tool in the past to solve various problems. In this research work, it is used to solve the problem of placing base stations in a given area in optimal way. Wireless communication has a great impact on our daily life and requires its working parameters to be optimum. These parameters include coverage area, hand off rates, number of calls per cell etc. All these and other parameters need to be optimally designed for a communication system to work optimally. Coverage area and interference are the two main parameters that are affected by the placement of base stations. If the base stations are placed very far from each other beyond their covering ranges, lot of spaces will be left with no coverage. On the other hand if they are placed very close to each other, interference will be the result. So the base stations need to be placed in optimum way. Placing the base stations requires work force which performs drive test to decide whether to place base station in a particular region or not. With our solution it only requires a computer to find the optimum placement of base stations. Given an area, our proposed methodology places the base stations in optimum way in that area. Neural network is used as a tool to find the solution to this problem.Item Microcomputer based methodology for distance protection on long uhv transmission lines using symmetrical components(UMT.Lahore, 2010) Mashood NasirThis research work proposes a methodology for the long UHV transmission lines distance protection using microprocessor for fault detection, isolation and auto reclosing processes. For the protection of transmission lines distance relays are used, which depend on a function of impedance between relay location and fault point. In the present era, the development of microprocessors, their extra efficient controlling and computing abilities can be utilized in distance relaying for efficient computing of fault distance on transmission line and the type of fault occurred. The approach that is adopted here is to use sequence components to calculate the distance of fault point. In a normal distance relaying system that uses phasor values of voltage and current as input, six impedance calculating units are required for all ten types of fault that may occur in transmission line. With the advent of symmetrical component analysis tool these six impedance calculating units can be replaced by a single unit which will overall result in a considerable optimized and efficient protection system. Using Symmetrical component theory a single performance equation is developed that will encounter all type of faults on transmission lines. Microprocessor will serve the purpose of computing sequence components, distance of the fault from the relay point and the type of fault. The phasor values of voltage and current is given as an input to the microcomputer that will calculate symmetrical components and resulting sequence impedances. It eventually finds the distance of the fault using performance equation, ratios of voltages in that particular equation and their resulting phase difference. These calculations will lead us to the judgment of the phases involved in fault. ETAP is used for simulation to obtain desired results. Although the formulation described here is independent of hardware yet it provides a complete analytical base for distance protection and is analyzed for different types of fault conditions using simulation tools.Item Optimization of n+1 queens problem using neural networks and ‘the proximity rule of initialization’(UMT.Lahore, 2010) Muhammad WaqasThe neural networks proved very handy in order to solve combinatorial optimization problems for the last two decades. Especially the Hopfield-Tank neural network model is extensively applied to obtaining an optimal/feasible solution to many different NP combinatory optimization problems like travelling salesman problem (TSP) and NP-hard combinatory optimization problem like N Queens Problem. This thesis describes a neural network optimizer/scheduler that optimizes a solution for a highly complicated version of N Queens Problem, i.e. N+1 non-threatening Queens on a N*N chessboard with an intermediate pawn on it. The behavior of the network is evaluated using asynchronous as well as synchronous mode of updating the neurons. Theoretical soundness of the network is established with simulation. Simulations show that the proposed neural network is capable of finding the optimized solution and is convergent to the global minima in 90% of the trials with polynomial average computational complexity.Item Efficient detection schemes for cooperative spectrum sensing in fading channels using cognitive radio networks(UMT.Lahore, 2010) Sadia MurshadAdvancements in wireless technologies are a breakthrough in revolutionizing the communication paradigm by providing ubiquitous internetwork access to large number of users at incredibly high data rates. The ever increasing deployment of mobile applications, networks and diverse services are a true depiction to this fact. As this trend continues the phenomenon of frequency spectrum occupancy is mounting up leading to spectrum scarcity. Analysis shows that for most periods the spectrum in control of majority primary users remains vacant and underutilized. This eventually implied to the idea of Dynamic spectrum allocation to users by Federal Communication Commission (FCC) which worked as a preventive measure against the issue of spectrum scarcity and underutilization. Cognitive Radios are sophisticated devices which can effectively sense the availability of frequency bands and can self modulate its parameters for dynamic transmission. In this thesis we have focused on Primary transmission detection techniques for spectrum sensing. Comprehensive analysis of energy detection, matched filter detection and cyclostationary feature detection is done with prime emphasis on their advantages and limitations. Based on these results we have proposed a two-step detection scheme which outperforms the existing models in terms of detection probability even at very low SNR value. The censoring capability of this detector leads to energy efficiency in cooperative environment. Analysis is done on Nakagami channel model for wireless communication which includes fading and multipath propagation effects. The system model discussed provides good agility for band shifting in case of interference or primary user presence because of reasonable sensing time which is very close to that of cyclostationary feature detector and can differentiate between desired signal and wideband noise.