Uploaded on

 

More in: Technology
  • Full Name Full Name Comment goes here.
    Are you sure you want to
    Your message goes here
    Be the first to comment
    Be the first to like this
No Downloads

Views

Total Views
258
On Slideshare
0
From Embeds
0
Number of Embeds
0

Actions

Shares
Downloads
2
Comments
0
Likes
0

Embeds 0

No embeds

Report content

Flagged as inappropriate Flag as inappropriate
Flag as inappropriate

Select your reason for flagging this presentation as inappropriate.

Cancel
    No notes for slide

Transcript

  • 1. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 FPGA IMPLEMENTATION OF FOUR PHASE CODE DESIGN USING MODIFIED GENETIC ALGORITHM (MGA) Kandarpa Srinivas, Bandi Sarada, Malijeddi MuraliAbstract— The proposed architecture consists of an I. INTRODUCTION TO SPREAD SPECTRUMefficient VLSI hardware implementation of theModified Genetic Algorithm for identifying the good Early spark-gap wireless transmitters actuallypulse compression sequences based on Discrimination used spread spectrum, since their RF bandwidthsFactor. The main advantage of implementation using were much wider than their information bandwidth.Hardware based Genetic Algorithm is its inherent The first intentional use of spread spectrum wasspeed over Software based methods. The speed probably used by Armstrong in the late1920s oradvantage makes the hardware based ModifiedGenetic Algorithm (MGA) is a prime candidate for early 1930s with wide-band frequency modulationreal time applications. This architecture provides the (FM). Global Positioning System (GPS) is now theflexibility of generating Pulse compression sequences world‟s single largest spread spectrum system. Thewith variable frequencies. Radar signal processing term „spread spectrum‟ describes a modulationapplications require a set of sequences with technique that makes the sacrifice of bandwidth inindividually peaky auto correlation and pair wise order to gain signal-to-noise (S/N) performance.cross correlation. Obtaining such sequences is a Because spread spectrum signals are noise-like,combinatorial problem. If the auto correlation and they are hard to detect. They are also hard tocross correlation are taken in the aperiodic sense then intercept or demodulate.there are hardly any theoretical aids available. Thusthe problem of signal design referred to above arechallenging problem for which many global II. PULSE COMPRESSIONoptimization algorithms like genetic algorithm,simulated annealing, tunneling algorithm were Pulse compression[2] techniques are usuallyreported in the literature. Higher performance applied in radar to increase the signal energysystems using custom silicon cost too much for the transmitted without sacrificing the range resolution.typically small production volumes, and are not It allows radar to utilize a long pulse to achieveflexible enough for research applications. Field large radiated energy but simultaneously to obtainprogrammable gate arrays offer the performance of range resolution of short pulse. The increasedcustom silicon while maintaining the economies andflexibility of the microprocessor based solutions. detection capability of a long pulse radar system isRecent FPGA devices possess the density and achieved while retaining the range resolutionperformance to realize Pulse compression sequences capability of a narrow pulse system.in a single FPGA. The Proposed VLSI architecture is The concept of pulse compression is shownimplemented on the FPGA as it provides the below figure where the short pulse, long pulse andflexibility of re-configurability and re- compressed pulse waveforms are shown FIG 1programmability. The Proposed architecture is anovel and efficient architecture as it generates the Short PulsePulse compression sequences at the clock rate ofFPGA.Index Terms—Discrimination Factor, FPGA, Long PulseModified Genetic Algorithm, Pulse CompressionSequences Manuscript received June, 2012 Compressed Pulse Kandarpa Srinivas, M.Tech VLSI 2’nd Year Student,(Electronics and Communications Engineering), AndhraUniversity/ Sanketika Vidya Parishad Engineering College, (e-mail: kandarpa.srinivas@gamil.com). Vizianagaram, India, Fig1 : Pulse Wave Forms+919985848985. Bandi Sarada, Assistant Professor,Electronics and Communications Engineering, Andhra Theoretically, in pulse compression, the codeUniversity/Sanketika Vidya Parishad Engineering College, is modulated onto the pulsed waveform duringVizianagaram, India, +919160695506., (e-mail:sarada.bandi@yahoo.co.in). transmission. At the receiver, the code is used to Malijeddi Murali, Professor & HOD, Electronics and combine the signal to achieve a high rangeCommunications Engineering, Andhra University/Sanketika resolution.The pulse compression receiverVidya Parishad Engineering College, Visakhapatnam, India, compresses the long received signal of length T+919440149970., (e-mail: muralitejas@gmail.com). 241 All Rights Reserved © 2012 IJARCET
  • 2. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012into a narrow signal of width l/B. It does this by The criteria of goodness of pulse-delaying each sub-part of the input signal compression sequences are evaluated by followingspectrum for different amounts so that each sub- methods.part arrives at the output at the same instant. E. Merit FactorThe pulse compression ratio is T/ (1/B) or B*T. The Merit factor (F) of a sequence is defined as the ratio of the main lobe energy of the III. FUNCTIONALITY OF SYSTEM autocorrelation to the total energy in the side lobes. Merit factor is a measure of the quality of pulseThis project consists of an efficient VLSI compression.implementation of Modified Genetic Algorithm, Let S = [x0, x1, x2, x3 ………..xN-1]for the identification of the good Pulse compression be a real sequence of length N.sequence based on Discrimination Factor. It is the n 1 kmain criteria for good Pulse compression sequence. r k   x x i ikThe architecture has been authored in VHDL for Let i 0four phase Pulse compression sequences and Where k=0, 1, 2… (N-1) is its aperiodicsynthesis was done with Xilinx XST. Xilinx ISE autocorrelation. Now Merit Factor is written asFoundation 9.1i has been used for performing r 2 0 mapping, placing and routing. For Behavioral F  N 1 2 r 2 k simulation and Place and route simulationModelsim6.0 has been used. The Synthesis tool k 1was configured to optimize for area and high effortconsiderations. The targeted device selected here is The merit factor F must be as large as possible forSpartan-3 xa3s5000fg900-5. The interest of the good sequence. The merit factor is an indication ofproject work is an attempt to obtain a real time the quality of pulse compression and a higher meritsignal processing VLSI architecture for the factor is always desirable.identification of the four phase pulse F. Discrimination Factorcompression sequences with good Discrimination The Discrimination factor is defined as thefor radar. Pulse compression sequences with low ratio of amplitude of main peak of the autoauto correlation functions with one main peak and correlation to the absolute maximum amplitude invery small side peaks are of interest in numerous the side lobes.practical applications in telecommunications, radar, r (0)and spread spectrum applications. D Maxk  0 | r k  | IV. PULSE COMPRESSION CODES AND ITS CHARACTERISTICS Discrimination Factor is used to measure weather a coded signal is Good or poor and also it measures A. Binary codes how the main lobe signal is different form the peak Binary is a way of counting using two side lobe level.numbers. Binary codes are used to represent G. Energy efficiencyinformation in digital form. It can be represented The energy efficiency is defined as theby (1,-1) or (0,1) or (0,-1). ratio of the actual energy in sequence to the energyB. Ternary Codes if every element has the maximum amplitude in the [9][6]Ternary Code is the code that can be sequence. Mathematically it is given as below, N 1 xused to represent information and data and uses 3 2digits for representation. (i.e.,) this code consists (k )of 1, 0, and -1. E k 0C. Four Phase codes N Four phase code can also be applied to For binary sequences the energyrepresent data. It uses four arguments to represent efficiency is unity. For ternary sequences it isinformation. Four phase code uses the alphabet -1, merely proportional to the non zero elements in the1, j, -j. sequence. of the For Four Phase sequences theD. Quinquenary codes Energy efficiency is better than ternary sequences. Quinquenary code uses five arguments to H. Quality Factorrepresent information. Quinquenary code uses the Quality factor is defined as the product ofalphabet -2, -1, +2, +1 and 0. merit factor and energy efficiency.E. Six Phase codes Q=F x E. Six Phase code uses six arguments torepresent information. Six Phase code uses thealphabet +1, -1, (0.5+j 0.866), (0.5-j 0.866), (-0.5+j0.866) and (-0.5-j 0.866) 242 All Rights Reserved © 2012 IJARCET
  • 3. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 V. INTIAL DESIGN IMPLEMENTATION STEPS A. Problem Solution A. Problem Formulation The problem is normally solved choosing the Problem formulation is the selection of appropriate techniques from those available in thedesign variables, constraints, objective function(s), field of optimization. These include gradient-basedand models of the discipline/design. algorithms, population-based algorithms, or others.B. Selection of design variables Very simple problems can sometimes be expressed Design variables can be continuous, linearly; in that case the techniques of lineardiscrete or Boolean. Design problems with programming are applicable.[3]continuous variables are normally solved moreeasily. Design variables are often bounded, that is, 1. GRADIENT-BASED METHODSthey have maximum and minimum values. A. Newtons MethodDepending on the adopted method, these bounds B. Steepest Descentcan be treated as constraints or separately. C. Conjugate GradientC. Selection of Constraints D. Sequential Quadratic Programming A constraint is a condition that must besatisfied to render the design to be feasible. An 2. POPULATION-BASED METHODSexample of a constraint in beam design is that the A. Genetic Algorithmsresistance offered by the beam at points of loading B. Particle Swarm Optimizationmust be equal to or greater than the weight ofstructural member and the load supported. In 3. OTHER METHODSaddition to physical laws, constraints can reflect A. Random Searchresource limitations, user requirements, or bounds B. Grid Searchon the validity of the analysis models. Constraints C. Simulated Annealingcan be used explicitly by the solution algorithm orcan be incorporated into the objective, by usingLagrange multipliers. B. Steps for the general procedure used toD. Objectives formulate and solve optimization problems. An objective is a numerical value that is tobe maximized or minimized. For example, a 1) Analyze the process itself to identify the processdesigner may wish to maximize profit or minimize variables and specific characteristics of interest,weight. Many solution methods work only with i.e., make a list of all the variables.single objectives. When using these methods, thedesigner normally weights the various objectives 2) Determine the criterion for optimization andand sums them to form a single objective. Other specify the objective function in terms of the abovemethods allow multi-objective optimization such as variables together with coefficients.the calculation of a Pareto front.E. Models 3) Develop via mathematical expressions a valid The designer has to also choose models to process model that relates the input-outputrelate the constraints and the objectives to the variables of the process and associated coefficients.design variables. These models are dependent on Include both equality and inequality constraints.the discipline involved. They may be empirical Use well known physical principles such as massmodels, such as a regression analysis of aircraft balances, energy balance, empirical relations,prices, theoretical models, such as from implicit concepts and external restrictions. Identifycomputational fluid dynamics, or reduced-order the independent and dependent variables to get themodels of either of these. In choosing the models number of degrees of freedom.the designer must trade-off fidelity with the timerequired for analysis. 4) If the problem formulation is too large in scope: break it up into manageable parts, or simplify the VI. FINAL DESIGN IMPLEMENTATION STEPS AND REPRESENTATION IN STANDARD FORM objective function and the model. Once the design variables, constraints, 5) Apply a suitable optimization technique forobjectives, and the relationships between them mathematical statement of the problem.have been chosen, Maximization problems can beconverted to minimization problems by multiplying 6) Examine the sensitivity of the result, to changesthe objective by -1. Constraints can be reversed in a in the values of the parameters in the problem andsimilar manner. Equality constraints can be the assumptions.replaced by two inequality constraints as follows. 243 All Rights Reserved © 2012 IJARCET
  • 4. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 most popular used EA technique, is explained as inC. Classical Optimization Techniques below figure 2. The classical optimization techniques areuseful in finding the optimum solution orunconstrained maxima or minima of continuousand differentiable functions. These are analyticalmethods and make use of differential calculus inlocating the optimum solution. The classicalmethods have limited scope in practicalapplications as some of them involve objectivefunctions which are not continuous and/ordifferentiable. These methods lead to a set ofnonlinear simultaneous equations that may be Fig 2 : A flowchart indicating the steps of a simpledifficult to solve. genetic algorithmD. Advanced Optimization Techniques[5] B. The pseudo code for above simple EA is given below A. Genetic algorithm B. Taboo search Algorithm i=0 C. Simulated Annealing Algorithm Initialize population P0 D. Hamming Scan Algorithm Evaluate initial population VII. GENETIC ALGORITHM while ( ! termination condition) [8]A genetic algorithm (GA) is a search {technique used in computer science to findapproximate solutions to optimization and search i = i+1problems. Specifically it falls into the category of Perform competitive selectionlocal search techniques and is therefore generallyan incomplete search. Genetic algorithms are a Create population Pi from Pi-1 by recombinationparticular class of evolutionary algorithms that use and mutationtechniques inspired by evolutionary biology suchas inheritance, mutation, selection, and crossover Evaluate population Pi }(also called recombination). The termination condition may be a desired fitness function, maximum number of generations etc.A. Evolutionary Algorithms for Optimization and After fitness function evaluation, individuals areSearch distinguished based on their quality. According to Most real world optimization problems Darwins evolution theory the best ones shouldinvolve complexities like discrete, continuous or survive and create new offspring for the nextmixed variables, multiple conflicting objectives, generation. There are many methods to select thenon-linearity, discontinuity and non-convex region. best chromosomes, for example roulette wheelThe search space (design space) may be so large selection, Boltzman selection, tournamentthat global optimum cannot be found in a selection, rank selection, steady state selection andreasonable time. The existing linear or nonlinear others. Two of these are briefly described, namely,methods may not be efficient or computationally roulette wheel selection and rank selection:inexpensive for solving such problems. Variousstochastic search methods like simulated annealing, Roulette wheel is a very common selectionevolutionary algorithms (EA) or hill climbing can method. Each individual is assigned a slice of abe used in such situations. EAs have the advantage circular “Roulette wheel” and the size of slice isof being applicable to any combination of proportional to the individual‟s fitness. The wheelcomplexities (multi-objective, non-linearity etc) is spun N times, where N is the number ofand also can be combined with any existing local individuals in the population. On each spin, thesearch or other methods. Various techniques which individual under the wheel‟s marker is selected tomake use of EA approach are Genetic Algorithms be in the pool of parents for the next generation.(GA), evolutionary programming, evolution Roulette wheel selection is easy to model and hasstrategy, learning classifier system etc. All these been used in most studies, but problems will ariseEA techniques operate mainly on a population when there are big differences between the fitnesssearch basis. In this project Genetic Algorithms, the values. For example, if the best chromosome fitness is 90% of the sum of all fitness, the other 244 All Rights Reserved © 2012 IJARCET
  • 5. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 chromosomes will have very few chances to be In this case both children are infeasible because selected. Parents are selected according to their they both contain repeated point. The best way to fitness i.e., each individual is selected with a handle the infeasible child strings is to use a probability proportional to its fitness value. In other different variant of crossover that does not allow words, depending on the percentage contribution to infeasible children to be created at all. the total population fitness, string is selected for mating to form the next generation. This way, weak 2. Mutation solutions are eliminated and strong solutions survive to form the next generation. Mutation occurs when a chromosome is changed in such a way as to alter the genetic message carried Rank Selection The individuals in the population by that chromosome. It is used to randomly alter are ranked according to fitness first, and then every the values of some of the positions in some of the chromosome receives fitness value determined by strings based on a parameter that determines the this ranking. The expected value of each individual level of the mutation. In our example, the second depends on its rank rather than on its absolute position in the string [1 -1 j -j] might be chosen for fitness. Its purpose is to prevent too-quick mutation and randomly switched from a value of -1 convergence. to j. This is an improvement: [1 -1 j –j] has fitness of 6.0 while this has a fitness of 11. Of course it is Random Selection A selection operator randomly just as possible that mutation could worsen the selects a chromosome from the population. fitness function or even generate an infeasible solution. The motivation behind the mutation is to B. Genetic Operators sample the solution space widely. So where crossover tries to concentrate the solutions that we 1. Crossover already have into better solutions, mutation works Crossover is a genetic operator used to vary the instead to sample the solution space and to broaden programming of a chromosome or chromosomes the search. Mutation is a vital part of the solution from one generation to the next. Crossover module process, and the mutation rate can have a big is used to perform the crossover operation on the 2 impact on the quality of the final solution. It is even winner individuals. During crossover two parent possible (though vastly more inefficient) to solve strings from the mating pool combine to create two the problems using only the mutation operator. new child solution strings. VIII. TABOO SEARCH ALGORITHM This happens as follows:1) Randomly select two parent strings from mating In most cases of practical interest, global pool. optimization is very difficult. This is because of the2) Randomly select a crossover point in the solution omnipresence of local minima, which tends to string. This is the point between any two positions increase exponentially with size of the problem. in the solution string. The principal requirement before any global3) Swap the ends of two parent strings, from the optimization method is that it must be able to avoid crossover point to end of the string to create two entrapment in local and continue the search to give new child strings. a near optimal solution whatever the initial The process is illustrated below. conditions are. The concept of the taboo search is based on a stochastic global optimization method Parent strings child strings originally developed by Glover for very large XXXX |XX XXXXOO combinational optimization tasks. The concepts of OOOO |OO OOOOXX “move” and “neighbourhood” are common to most Crossover point heuristic and algorithmic procedures. In taboo search, a move is a transition from one trial solution to another. The move value is the Here X and O represent values in the solution difference of trial solutions. The move improves strings. There are numerous variations in the basic only if the move value is negative. In order to avoid crossover operator, for example randomly choosing a blind search, taboo search uses a prescribed two crossover points and swapping the string problem specific set of contents, known as “taboo contents between those two crossover points. Of conditions” which must be satisfied for the move to course it is possible that crossover may result in be considered admissible. Otherwise, the move is infeasible children, as for example: taboo. A move remains taboo only during the “taboo period” a certain specified number of We might see crossover such as iterations. The “aspiration conditions” is defined to CDA|B CDAC enable certain “interesting” moves. If this is BAD|C BADB satisfied, a taboo move becomes admissible. The 245 All Rights Reserved © 2012 IJARCET
  • 6. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012taboo condition and the aspiration condition fitness is evaluated, if the fitness value hastogether are a heuristic device, a kind of learning improvement over the original sequence the changeprocedure that benefits from information acquired is accepted otherwise the original sequence isduring previous iterations. Thus taboo search retained. The same procedure is performed for allperforms an intelligent search of the solution space. the elements of sequence. This process isFrom among the admissible moves at each iterative recursively applied until no changes are required. Astep, taboo search accepts the move with the lowest single mutation in a sequence results in a hammingmove values. This move might not lead to better distance of one from the original sequence. Thesolutions, but enables the algorithm to continue the Hamming scan algorithm mutates all the elementssearch without becoming confounded by the in the sequence one by one and looks at all theabsence of improving moves and to “climb out” of first order hamming neighbors of the givenlocal minima. sequence. Thus hamming scan performs recursively local search among all the hamming-1 IX. SIMULATED ANNEALING ALGORITHM (SAA) neighbors of the sequence and selects the one whose objective function value is minimum. The [10][7][4]The name and inspiration come search time required for Hamming scan algorithmfrom annealing process in metallurgy, a technique increases very fast with the length of sequence asinvolving heating and controlled cooling of a the algorithm performs recursive search among thematerial to increase the size of its crystals and Hamming neighbors of the element in thereduce their defects. The heat causes the atoms to sequence. Thus, Hamming scan also becomesbecome unstuck from their initial positions (a local unaffordable at larger lengths. The Hamming scanminimum of the internal energy) and wander is expedited and hence made applicable at largerrandomly through states of higher energy; the slow lengths.cooling gives them more chances of findingconfigurations with lower internal energy than the XI. MODIFIED GENETIC ALGORITHM (MGA)initial one. Simulated Annealing (statisticalcooling) performs a computation that is analogous Modified Genetic Algorithm is proposed as ato a physical process. In the process a material is statistical technique for obtaining approximatefirst heated up to a temperature that allows all its solutions to combinatorial optimization problems.molecules to move freely around and is then cooled The proposed algorithm is a combination ofdown slowly. The freedom of movement for the Genetic algorithm (GA) and Hamming scanmolecules decreases gradually until all the algorithm. MGA combines the good methodologiesmolecules take a fixed position. At the end of the of two algorithms like global minimum convergingprocess, the total energy is minimal provided that property of GA algorithm and fast convergence ratethe cooling is very slow. The energy corresponds to of Hamming scan algorithm. The demerit ofthe cost function. The movement of molecules Hamming scan algorithm is that it gets stuck in thecorresponds to a sequence of moves in the set of local minimum point because it has no way tofeasible solutions. The temperature corresponds to distinguish between local minimum point and aa control parameter T. In the simulated annealing global minimum point. Hence it is sub-optimal.method, each point of the search space is compared The drawback in genetic algorithm is that it has ato a state of some physical system, and the function slow convergence rate because even though it mayto be minimized is interpreted as the internal get closer to the global minimum point, it may skipenergy of the system in that state. Therefore the it because of the methodology it employs. Thegoal is to bring the system, from an arbitrary initial MGA overcomes these drawbacks. Geneticstate, to a state with the minimum possible energy. algorithm (GA) is a class of search algorithm that have been used in many optimization problems. X. HAMMING SCAN ALGORITHM GA have an analogy with biological evolution. This evolution leads to efficient exchange of The Hamming scan algorithm is a information between all models encountered, andtraditional greedy optimization algorithm, which allows the algorithm to rapidly assimilate andsearches in the neighborhood of the point in all exploit the information gained to find better datadirections to reduce the cost function and has fast fitting models. GA inherently superior to randomconvergence rate. The basic difference between search techniques and can also perform better thanGenetic algorithm and Hamming scan algorithm is iterative matrix inversion which requires a goodthat Genetic algorithm uses random but possibly starting models The important ingredient of themultiple mutations. This algorithm mutates Genetic approach is that large and complex modelselements of sequence one by one. The mutation is a are represented by binary strings. These bit stringsterm metaphorically used for a change in an can be manipulated to produce more successfulelement in the sequence. For example, consider a information by two main GA operations Crossoverfour-phase sequence [1 -1 j j] whose fitness is and Mutationevaluated. The value 1 is replaced with –j and 246 All Rights Reserved © 2012 IJARCET
  • 7. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012 start XII. FOUR PHASE CODE DESIGN USING MODIFIED GENETIC ALGORITHM Since Discrimination Factor is the main criterion Initialize the variables L,N. for good pulse compression sequences, therefore the Four Phase sequence having minimum side lobe amplitude can be considered as the best Four Compute the Fitness Function Phase Pulse compression sequence. The Block (Discrimination Factor) of all the randomly generated Diagram of a Modified Genetic Algorithm for the sequences. Design of Four Phase pulse Compression Sequence is shown below figure 4. FIG 4 : BLOCK DIAGRAM OF MGA Crossover Module. Random Sequence Generation Compute the Fitness Function. Fitness Function Mutation Module. Compute the Fitness Function. Crossover Module MModule Hamming Scan. Mutation Module Compute the Fitness Function Hamming Scan Module No Is Fitness Function Fitness Function value greater than optimum value? Min. Optimal Yes Amplitude Sequence Stop XIII. RESULTS AND ANALYSISFig 3 : MGA Flow Chart A. Simulation Waveforms Four phase sequences are designed usingThe three main steps involved in GA are the proposed novel and efficient VLSI architecture. The synthesized results are single realizations1First the initial sequence is generated. obtained using Spartan-3. The behavioral simulation waveforms for2The evaluation values of the fitness function for the good Four Phase Pulse compression sequencethe current sequence is calculated. of length 4 are shown in Fig.5.3. The termination criteria is checked. The wholeGA procedure stops if the termination criteria isreached otherwise selection ,crossover , mutationand Hamming scan operations are performed. After the synthesis ,the synthesis tool generates a complete net-list of target hardware components 247 All Rights Reserved © 2012 IJARCET
  • 8. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012and their timings. The details of gates used for the The behavioral simulation waveforms forimplementation of the design are described in this the good Fourphase Pulse compression sequencenetlist of length 20 are shown in Fig10. It can be seen that The Post-Route simulation waveforms for the Fourphase sequence based on the lowest sidelobegood Four phase Pulse compression sequence of amplitude is 10 among all the sidelobes is 001 101length 4 are shown in Fig.6 111 011 101 111 011 101 011 101 011 101 011 111 111 001 101 111 001 001(1 j –j -1 j –j -1 j -1 j -1 j - 1 –j –j 1 j –j 1 1). Therefore Discrimination Factor of this sequence is 6.3245The netlist also includes wiring delays and loadeffects on gates used in the post-synthesis design.The netlist output is made available in variousnetlist formats including VHDL. Such a descriptioncan be simulated and its simulation is referred asPost Synthesis simulation. Timing issue ,Determination of proper clock Frequency and race The behavioral simulation waveforms for the goodand hazard considerations can only be checked by a Fourphase Pulse compression sequence of lengthpost synthesis simulation run after a design is 32 are shown in Fig.11. It can be seen thatsynthesized. Fourphase sequence based on the lowest sidelobe amplitude is 26 among all the sidelobe amplitudesThe behavioral simulation waveforms for the good is 001 011 011 001 101 001 101 111 001 001 101Four phase Pulse compression sequence of length 8 011 111 101 011 011 001 011 111 001 111 011 101is shown in Fig.7 111 001 111 111 001 111 111 011 001 ( 1 -1 -1 1 j 1 j –j 1 1 j -1 –j j -1 -1 1 -1 –j 1 –j -1 j –j 1 –j –j 1 –j –j -1 1). Therefore Discrimination Factor of this sequence is 6.2757. The Post-Route simulation waveforms forthe good Four phase Pulse compression sequenceof length 8 are shown in Fig.8 B. FPGA Editor Routed Design The FPGA Editor is a graphical application for displaying and configuring Field Programmable Gate Arrays (FPGAs). The FPGA Editor requires a Native Circuit Description (.ncd) file. This file contains the logic of our designThe behavioral simulation waveforms for the good mapped to components (such as CLBs and IOBs).Fourphase Pulse compression sequence of length In addition, the FPGA Editor reads from and writes16 are shown in Fig.9 It can be seen that Fourphase to a Physical Constraints File (PCF).The followingsequence based on the lowest sidelobe amplitude is is a list of a few of the functions you can perform2 among all the sidelobes is 111 011 101 001 011 on your designs in the FPGA Editor.101 101 101 001 101 111 101 0111 011 111 001(-j-1 j 1 -1 j j j 1 j –j j -1 -1 –j 1). Therefore 1 Place and route critical componentsDiscrimination Factor of this sequence is 11.3137. before running the automatic place and route tools. 2 Finish placement and routing if the routing program does not completely route your design. 3 Add probes to your design to examine the signal states of the targeted device. Probes are 248 All Rights Reserved © 2012 IJARCET
  • 9. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 2012used to route the value of internal nets to an IOB XII. CONCLUSIONfor analysis during the debugging of a device. 4 Run the BitGen program and download The proposed VLSI System is a unique real-the resulting BIT file to the targeted device. time signal processing solution for synthesis of 5 View and change the nets connected to optimal Four phase codes which has the followingthe capture units of an ILA core in your design. advantages. It has high efficiency as it identifies 6 Use the ILA command to write a .cdc the optimal sequences within a short time, butfile exhaustive search consumes more time as the 7 Create an entire design by hand length of the sequence increases. The proposed(advanced users). VLSI System is a single chip solution for both 7 Discrimination Factor and Merit Factor identification and generation of pulse compressionof Fourphase pulse Compression Codes sequences with a little addition to the proposed The Fourphase sequences have superior architecture. The most computational intensive partDiscrimination factors when compared to the merit of MGA is the fitness evaluation, each individualfactors of Fourphase sequences. These can be population evaluates efficiently in MGA. Theserepresented in the form of a graph as shown in Fig improved FPGA technologies could be exploited to5.12. Here blue color line represents improve the MGA s capabilities. The parallelismDiscrimination factor and pink color line through the FPGA pipelining contributed therepresents Merit factor. Sequence lengths on X-axis majority of the hardware‟s speed up in this, Discrimination and Merit factors are indicated on application. The concurrent execution of thethe Y-axis. Sequence lengths ,Discrimination FPGA, the hardware memory access efficiency,Factor values and the Merit Factor values are and its efficiency in generating random numbersshown in Table 12 are also noticeable factors in identifying four phase Table 13 Graphical representation of pulse compression codes. The synthesized Fourcomparing Discrimination Factor and Merit Factor phase sequences have good Discrimination Factorof fourphase codes using MGA. and less cross correlation energy values. Hence they are promising for practical applications such as Spread Spectrum and Multimedia Applications. It was also observed that the proposed architectureTable 12 is giving good Discrimination Factor values even for higher sequence lengths. This shows the COMPARISION OF DF WITH MF superiority of the architecture. 12 Discrimination Factor 10 REFERENCES 8 Discrimination Factor 6 Merit Factor 1. Ackroyd, M.H., “Synthesis of Efficient 4 Huffman Sequences”. IEEE Trans. On AES- 2 8.pp.2-8, Jan.1972. 0 2. Ackroyd, M.H., “Huffman sequences with 8 10 12 16 20 28 30 33 35 40 46 50 Sequence Length appropriately uniform envelopes or cross correlation functions”, IEEE Trans. On IT, vol.23,pp.620-623,Sept.1997Table 13 3. Brewer. J.W, “Kronecker products and matrix calculus n system theory”, IEEE Trans. On CAS, Vol.25, No9, pp. 772-781,Sequence Discrimination Merit 1978. Length Factor Factor 4. Brown, D. E., Simulated Annealing. In 8 6.364 5.33 Linear Programming, J. P. Ignizio and T. M. 10 7.0711 5.55 Cavalier. New Jersey: Prentice-Hall. 12 8.4853 7.2 5. K. Deb. “Optimization for Engineering 16 11.3137 9.142 design: Algorithms and examples”, Delhi: 20 7.0711 5.556 Prentice-Hall,1995 28 8.8544 5.9394 6. Gavish. A & Lempel .A., “On Ternary 30 10 4.94 complementary sequences”, IEEE Trans. on 33 10.4355 5.0417 IT, vol.40,no.2,pp.522-526,March 1994 35 9.7073 3.5819 7. Kirkpatrick .S, C. D. Gelatt, and M. P. 40 9.4281 4.7059 Vecchi, "Optimization by Simulated 46 10.2859 3.7385 Annealing," Science, vol. 220, pp. 671-680, 50 9.8058 3.7994 May1983. 249 All Rights Reserved © 2012 IJARCET
  • 10. ISSN: 2278 – 1323 International Journal of Advanced Research in Computer Engineering & Technology Volume 1, Issue 5, July 20128. Maryam Amin Nasrabadi “A new approach Design Integr. Circuits Syst., vol. 9, no. 7, for long low autocorrelation Binary pp. 767– 778, Jul. 1990. sequence problem using Genetic algorithm 22. Pei-Yin Chen,Ren-Der Chen,Yu-Pin “IEEE proceeding, 2006. Chang,”Hardware Implementation For a9. Moharir. P.s., “Ternary Barker codes”, GeneticAlgorithm”, IEEE Trans. Electron Lett, vol.10, pp 460-461, 1974. Instrumentation and Measurement,vol.10. Moharir. P.s., Varma. S.K & venkat rao.k. 57,NO.4,april 2008 “Ternary pulse compression sequences”, 23. Xilinx, Spartan-3 Field Programmable Gate J.IETE, vol.31, pp.33-40, 1985 array data sheets.11. Moharir P.S., Singh .R & Maru V.M., “S-K- 24. The data taken from H algorithm for signal design”, Electronic http://en.wikipedia.org/wiki/simulated Letters, Vol.32, No.18, pp. 1648-1649, anneling August 1996.12. Nathonson, F.E., “Radar design principles”, McGraw-Hill Book Co., New York. 1969.13. Rihaczek.R.W, “Principles of high resolution radar”, McGraw-Hill, New York., 1969. Kandarpa Srinivas received the B.Sc in14. Skolnik.M.I. “Introduction to Radar Electronics and the M. Sc Electronics from Systems”, McGraw Hill International the Andhra University (AU), Visakhapatnam, in 2005 and 2007 edition, second edition, pp.399-438, 1980. respectively. Currently working toward the15. Woodward, P.M., “Probability and M.Tech degree in VLSI. Worked as information theory, with application to Lecturer in TP Polytechnic College between radar”, pergamon press, oxford, 1953. June 2008 and May 2010.Fields of intrest is VLSI and Communication Systems.16. N. Balaji, Dr .K. Subba Rao, M. Srinivasa Rao, “Hybrid GA-SA algorithm for Non Binary Pulse Compression Sequence” Proc Bandi Sarda received the B.E in ECE and of National Conference on Communication the M.E Electronics &Instrumentation from Andhra University (AU).Currently working & Signal Processing April 12-13, 2007. as Assistant Professor in Sanketika Vidya17. N. Balaji, Dr .K. Subba Rao, M. Srinivasa Parishad Engineering College. Rao, “FPGA Implementation of the Ternary Pulse Compression Sequences” Proc of Malijeedi Murali received the B.Tech in ECE and M.E from IAENG International Conference on Andhra University. He also received MBA degree. Currently Computer Science (ICCS08 19th -21st working toward the PhD in Communications Systems in Andhra March 2008. University. He research has published in leading 10 National18. N. Balaji, Dr. K. Subba Rao, M. Srinivasa and 5 International Journals and in Conferences. Rao, “FPGA Implementation of Four phase Pulse Compression Sequences using FPGA” Proc of Third Innovative Conference on Embedded Systems, Mobile. Communication and Computing (ICEMC2) 11th - 14th August, 2008.19. Singh. S.P, N. Balaji, Dr. K. Subba Rao, “Sixty-phase Code Design Using Modified GA”, Proc of International Conference on Recent Advances in Communication Engineering (RACE-08), 20-23 December 2008.20. N. Balaji, Dr .K. Subba Rao, M. Srinivasa Rao, “FPGA Implementation of the Binary Pulse Compression Sequences with Good Merit factor” ,Proc of International Conference on Computing, Communication and Control (ICAC3‟09)January 23-24, 2009.21. M. Serra, T. Slater, J. C. Muzio, and D. M. Miller, “The analysis of One dimensional linear cellular automata and their aliasing properties,” IEEE Trans. Comput.-Aided 250 All Rights Reserved © 2012 IJARCET