NATIONAL CONFERENCE ON ELECTRICAL AND COMPUTER ENGINEERING | Year:2016 | دوره:3 |Papers Count:5

AN EXCELLENT CANDIDATE FOR A HIGH-QUALITY PORTABLE FREQUENCY STANDARD AT THE TELECOMMUNICATION REGION IS THE 5S1.2 (F=3) –5D5.2 (F=5) TWO-PHOTON TRANSITION OF RUBIDIUM (85) AT 778 NM. THE INTERROGATION OF THE TWOPHOTON TRANSITION NEEDS EFFICIENT SECOND HARMONIC CONVERSION OF A LASER AT 1556 NM. IN THE PRESENT WORK, A COMPARISON IS MADE BETWEEN THE CONVERSION EFFICIENCIES OF BULK AND WAVEGUIDE PERIODICALLY POLLED LITHIUM NIOBATE (PPLN) CRYSTALS. IN ORDER TO ACQUIRE THE OPTIMUM OPERATING CONDITIONS, THE FREQUENCY DOUBLED POWER IS MEASURED AS A FUNCTION OF CRYSTAL TEMPERATURE AND LASER WAVELENGTH. THE CONVERSION EFFICIENCY OF THE BULK PPLN CRYSTAL IS FOUND TO BE SMALLER THAN THAT OF THE WAVEGUIDE PPLN FOR INPUT POWERS LESS THAN 150 MW.

AN EXCELLENT CANDIDATE FOR A HIGH-QUALITY PORTABLE FREQUENCY STANDARD AT THE TELECOMMUNICATION REGION IS THE 5S1.2 (F=3) –5D5.2 (F=5) TWO-PHOTON TRANSITION OF RUBIDIUM (85) AT 778 NM. THE INTERROGATION OF THE TWO-PHOTON TRANSITION NEEDS EFFICIENT SECOND HARMONIC CONVERSION OF A LASER AT 1556 NM. IN THE PRESENT WORK, A COMPARISON IS MADE BETWEEN THE CONVERSION EFFICIENCIES OF BULK AND WAVEGUIDE PERIODICALLY POLLED LITHIUM NIOBATE (PPLN) CRYSTALS. IN ORDER TO ACQUIRE THE OPTIMUM OPERATING CONDITIONS, THE FREQUENCY DOUBLED POWER IS MEASURED AS A FUNCTION OF CRYSTAL TEMPERATURE AND LASER WAVELENGTH. THE CONVERSION EFFICIENCY OF THE BULK PPLN CRYSTAL IS FOUND TO BE SMALLER THAN THAT OF THE WAVEGUIDE PPLN FOR INPUT POWERS LESS THAN 150 MW.

RECENTLY, META-HEURISTIC OPTIMIZATION ALGORITHMS ARE USED TO FIND OPTIMAL SOLUTIONS IN HUGE SEARCH SPACES. ONE OF THE MOST RECENT IS IMPERIALIST COMPETITIVE ALGORITHM (ICA) WHICH IS WIDELY USED IN MANY OPTIMIZATION PROBLEMS AND HAS SUCCESSFUL RESULTS. WE ADD SOME ELITISM TO ICA AND INTRODUCED ELITIST IMPERIALIST COMPETITIVE ALGORITHM (EICA) AS A NEW VERSION OF ICA.ONE OF THE MOST IMPORTANT APPLICATION OF OPTIMIZATION TECHNIQUES IS IN DATA MINING WHERE CLUSTERING AND ITS MOST POPULAR ALGORITHM, K-MEANS, IS A CHALLENGING PROBLEM. ITS PERFORMANCE DEPENDS ON THE INITIAL STATE OF CENTROID AND MAY TRAP IN LOCAL OPTIMA. IT IS SHOWN THAT THE COMBINATION OF EICA AND K-MEANS HAVE BETTER PERFORMANCE IN TERMS OF CLUSTERING AND EXPERIMENTAL RESULTS ARE DISCUSSED ON K-MEANS CLUSTERING. THE GOAL OF THIS RESEARCH IS TO IMPROVE ICA FOR ANY OPTIMIZATION PROBLEM.

A WIDEBAND EDGE-FED MONOPOLE ANTENNA WITH ARTIFICIAL MAGNETIC CONDUCTOR (AMC) AS GROUND PLANE IS PROPOSED. AMC GROUND PLANE INSTEAD OF CONVENTIONAL PEC PLANE IS USED TO IMPROVE THE ANTENNA GAIN AND BANDWIDTH. THE PROPOSED AMC STRUCTURE WHICH USES A 1 MM AIR GAP RESULTS IN INCREASING THE BANDWIDTH TO ABOUT 18.25% WITH NEARLY 4% IMPROVEMENT COMPARED TO THE AMC STRUCTURE WITHOUT AIR GAP. USING AIR GAP AND REDUCING AMC SUBSTRATE HEIGHT IS A SUITABLE METHOD FOR DECREASING WEIGHT, ESPECIALLY IN LOWER FREQUENCIES. IN THE CASE OF USING AIR GAP A 1-MM-THICK AMC SUBSTRATE CAN BE USED THAT IMPROVES BANDWIDTH TO 23.91% AT THE EXPENSE OF ONLY 1DB REDUCTION OF THE GAIN WHICH IS A DESIRED RESULT.

CONVOLUTIONAL NEURAL NETWORK HAS GAINED ENORMOUS SUCCESS IN RECENT YEARS, AND IS ONE OF THE MOST POPULAR DEEP LEARNING ALGORITHMS THAT HAS BEEN EXTENSIVELY USED IN MANY MACHINE LEARNING RELATED FIELDS. THE SUCCESS AND DIFFERENT APPLICATIONS OF CNN HAVE BEEN STUDIED AND ADDRESSED IN MANY STUDIES IN THE LITERATURE, HOWEVER, SOME ASPECTS WHICH INTERESTINGLY ARE VERY IMPORTANT ARE EITHER LESS WORKED ON OR IGNORED COMPLETELY. IN THIS PAPER WE STUDY AND ADDRESS SOME OF THE ASPECTS AND RESPECTIVE TRENDS THAT AFFECT THE APPLICATION OF CNN IN VARIOUS FIELDS.