Document Type: 
Document Number (FOIA) /ESDN (CREST): 
Release Decision: 
Original Classification: 
Document Page Count: 
Document Creation Date: 
November 2, 2016
Document Release Date: 
September 17, 2001
Sequence Number: 
Case Number: 
Publication Date: 
December 31, 1967
Content Type: 
PDF icon CIA-RDP86-00513R000721820011-9.pdf4.53 MB
KI 1 AK"17 VI C11 fA PA 1h5T93 UM/Physics -'Resonators Aug 51 "Analysis of Impulses by Neaw, of Resonators," A. A. Kharkevich "Zhur Tekh Fiz" Vol XXI, No 8i pp 886-891 Subject analyzers were applied to periodic pheno- mena. Recently they vere applied to nonperiodic phenomena, like Impulses. Author shoos that unA r certain conditions resonators may give vith- out extinction discrete values of sp:-ctral density of an impulse. Submitted 31 Jan 51. I*.-JM,vTcTT) A Spektry i Analiz. !zd. 2.J. Ispr. i Dop. Moskva, Gostekhizdat, 1953. 215 p. Diagrs., graphs bibliography: p. 21h-215. Nil C-It PITASE II TREASURE ISLAND BIBLIOGRAPHICAL REPORT AID 249 - ii BOOK Call No.: AF594622 Author., MIARKEVICII, A. A. Full Title: SELF-EXCITED OSCILLATIONS Transliterated Title: Avtokolebaniya Publishing Data Originating Agency: None Publishing House; State Publishing House of Technical and Theoretical Literature Date: 1953 No. pp.: 170 No. of copies: 10,000 Editorial Staff: None Text Data Coverage: This book describes physical anpe~cts of' self-excited osell- lationz. The study is.conducted without mathematical analysis and Is based on energy correlations. A large number of examples of self- excIted oscillations in physics and engineering has been analysed. Preface: The leading role of the Soviet science in the field of non- linear oscillation, and particularly in self-excited oscillations, is generally well known. Books like Theory of Oscillations, by A. A, Andronov and S. E. Khaykin, Self-Excited Oscillating yLterps, by K, F. Teodorchik and others are excellent and exceptional In thls field. The present small book may prove to be of value because it *hsr USSR/Electronics Information Theory I- ji i~~ 't'_ 1~~ k,- ~, k -A - Card 'I P"u'b. 'go-1/14 Author Kharkevich, A. A. Title Main outline of the general communication theory Periodical : Radiotekhnika 9, 3-7, Sep/Oct 1954 Abstract : The author outlines the development of the communication or information theory,*starting with Hartley's work (1928) on the notion of the con- tent of a message, down through Shannon's formulation of the comprehen. sive theory. He discusses V. A. Kotellnikov's theorem (1933) enabling the translation of any type of message (cpntinuous or in discrete unitb) into the transmission of discrete numbers at a definite rate, the com- promise between economy of units and the need for error corrections the correlation method of reception, D. V. Ageyev's theorem (1935) of signal separation, and problems of methods. He states that Soviet scientists should reject "cybernetic philosophy" but not "e~bernetic techniques." Four references: 3 USSR (1933, 1935, 1946). (Article is the author's abstract of a paper he delivered at a city-wide meeting of membcrs of the All-Union Scientific and Technical Society of Radio Engineering and Electric Communications imeni A. S. Popov [VNORiE] at Moscow on April 26, 1954). 3~ns'titution Submitted ~r4i ~ : 1, .':. -ot. ;': ~ .1 * I --*.i ` ~ .1 - . .- -Y, . KAZARTAN, Ra&~11 Avetisovich; VNSHIKOV, Boris Ivanovich; INARIVICH A#A#, redaktor; ANDFJff=O,Z.D,,, redaktor; KHEMSKAY TeMn-Icheekly redaktor [Transmission of messages through the cowunication system) Pe- redwha soobahchonii po sistemax eviazi. Moskva, Goo.izd-vo lit- ry po voprosam evissi i radio, 1955. 41 p. (MMA 9:2) (Telecommumication) KILUUETVIN Aleksandr Alsksandrovich; KARAS%7, M.D. redaktor; MURA- radaktor (Outline of a general theory of communication] Ocherki obshchei teorii aviazi. Moskva. Goa. izd-vo tekhniko-teoret. lit-ry. 1955. 268 p. (MLRA 8:8) (Information theory) UsSR/Electronic3 Information Theory FD-2928 Card 1/1 Pub. 41-9/17 Author :Blokh, E. L. and Kharkevich, A. A., Moscow Title :Geometric presentations in the theory of communications Periodical :Izv. AN SSSR, Otd. Tekh. Nauk 6, 91-100, June 1955 Abstract :Describes how and where geometric theory can be applied to the study and the science of communications. Discusses the inci- dental vector, the message and the signal, the signal and inter- ference, interference rejection, transmission capacity, maximum transmisaion capacity, method of storage and the separation of signals. Diagrams, formulae. Seven references, 4 USSR. Institution Submitted April 14, 1955 USS M/Electronics luformatina Rrzory FD-266c', Card 1A Pub~90-1/i2 Author Blokh, E. 1w. and Kharkev'-ch, A. A. Title Geometric theory of the threshold of transmission capacity of a coimnunicati-,)ns system Periodical Radiotekhnika., 10, 3-7, Jul 55 Abstract 'The limiting factors of signal transm:lGsion 1i a cormunicatlQns system are -evaluated on the basis of the geumf-tric theory. The limiting transmissicn capacity of a system is d,~fiaed as the greatest a=unt of in!'elligence that can be conveyed to the re- cc-.ivin6 end of the 1-Ine, maintaining tha lowest desired prob- ability of error. The transmission capacity approaches zero as the level of noise approaches that of the signal. Reliable re- ception for small incremerts of signal over noise require the use of sFeeial method3 of reception, such as storage and correla- tion methods. Graphs. Two references; one USM3. Institution Submitted January 11, 1955 KHMIN, DO- m6peatm and anAl"IS." A. A. Markovich. Roviewed bY IA. Mmrgin. Urp,mat.nauk. 10 no.l:Z3Y-z4z 'DD (MBA 8:6) (Splotr= analysts)(KathanatiOl physics) -r~i Rk -'C :ation Theory FD - 1933 Card 1/1 Pub 90-2/9 Author Kharkevich, A. A., and Blokh, E. L. Title a,;i~ingcapacity of a communication system PeAodical Radiotekhnika 10, 14-20, Feb 1955 Abstract The derivation of an expression based on geometrical relationships for determining the limiting capacity of a communication system is given. The older, well-known Shannon formula generally used for these calcu- lations holds true only when the silgnal-to-noise ratio approaches infinity. Institution: Submitted December 15, 1954 -.- ,- Ir j)), K j -~,-,-- K BLOIR, B.L.;KHARKNVICH, A.A. Reply to L.M.Fink's remarkL Radiotekhnika 10 no.10:75 0 '55- (Telecommunication) (HLRA 9:1) MAR VICH. Alelcoandr Alakeandrovich- XOSTIM&O.A.Z., redaktor; ~ a TUMAIUUIVL#---H-; ., e"MuluTMTKIv redaktor [Notilinear and parametric phenomena in radio enegineering] lelineinye I paramet riche skie iavleatia v radiotekhnike. M013kV9. 006. 12d-vo tekhaiko-teoret. lit-z7, 1956. 184 p, (MLRA 10:1) (Radio aircuito) K,'AIKFVICI*) A. A., (Prof.) "Questions of the Theory of Information in Systems of Automatic Recuation, Telecontrol and Telemetering," pappr rcpd at tl~e Session of the Acad. Sci. USSR, on Scientific Problems of Automatic Prodliction, 15-20 October 1956. Avtomatika i telemekhanika, No. 2, p. 182-192, 1957. 1 9015229 KHAR VICH, A.A., doktor tokhnichookikh ujudc. On the best cod*. Blektrosylaz' 10 no,2:65-70 .7 156.(KERA 9:6) (Ciphor and telograPh codes) KHARK3VICH,,,&,A- On the theory of an ideal receiver. Slaktrosviaz' 10 no.4:28-34 Ap 156. (Radio--Raceivers and reception) (MLRA 9:7) KV1WV-V'Cd, SUBJECT USSR / PHYSICS CARD 3 PA - 1705 AUTHOR BLOH,E.L., RARKEVIC,A.A. NONIS Theorem. TITLE On the Question of the Geometric Proof of SHAN PERIODICAL Radiotechnikaq 11, fasc. 119 5-16 (1956) issued: 12 / 1956 In the course of previous works (Radiotechnika, fasc.2 and 7, 1955) the authors endeavored to prove the theoren, on the pezietrabilit~ limit geometri- cally. Accordinf; to SHANNON this theorem is: C - F loG + Pn 2 P1) P here denotes the average power of the transmitter, P - the power of the perturbation in the stripe F, C - velocity. lit the present work the theorem is presented in MANNON'S form and also geometric proof of the second state- ment made in this theorem. It was found that SHANNION failed to take the follow- Ing into account: 13yun in thu case of the densest arrangement the coefficient of the filling up of the snace by non-intersecting sphere8 is diminished If n = 2FT (T - time, n - dimension) increases, and at n -~ oo it tends towards zero. The authors corrected this error committed by SHANNON and obtained an expression which deviates from that of SHANNON: C -* F o g ( 1 + L-) - 11 Ll Pn The difference between the two formulae is very essential in the case of com- parable P and Pn , namely just in the case of such conditions as are of par- ticular interest lit modern radiotechnology. On the other hand, SHANNONIS formula has been generally accepted. This contradiction could be explained by Radiotechnika, i1, faso. 11, 5-16 (1956) CARD 3 / 3 PA - 1705 authors regret not being in possession of this proof. It would be of im- portance because then not only SHAjnjOjj9S theorem could have been proved eecmetrically, but it would have been possible to show whether the limit of penetrability can be reall-zed by means of a receiver that is ideal in KOTELJNIKOV'S sense. INSTITUTION; -sandr Ale .KHARKIVICH. A le-k .AUqAUv"WKOSTIYANKO. A.1., red.; GAVRILOV, S.S.. 70 kwn-. Fe- Trw""~ (Spectra anct analysis] Siektry i analis. IzI. 3-e. parer. Moskva, Go~. izd-vo tekhniko-teoret. lit-x7, .1957. 236 p. (HIRA 11:2) (Speotrum analycia) KAZARYAII. Rafael' Avetisovich; KINSHINOV, Boris Ivanovich; IIAZAROY. Mikhail Vas illyevi ch, BERG, A.I.,rodaktor; D*IGIT, I.S.,redaktor; KULIKOVSKIY, A.A.,rad*ktor; SMIRMOV, A.D.,radaktar; TARASOV, F.I.,rodaktor; TRAHM. B.F..redaktor'; CHICHIK, P.O.,redaktbr; SHAKSHUR, V.I.,redaktor; ,"A.A.,redaktor; FMDVBDBY, L. Ya., %akhniches~iy.redaktor * _p. ~.Ozu - - [Xlembnts of the general theory of communications) Xlementy obahchei teorii sviazi. Koskya. Goo. energ. izd-yo. 1957- 94 p. (Hsasovaia radiobibliotaka. no.263) (K[aA 1o:4) (Telecommonication) PHASE I BOOK EXPLOITATION 476 Kharkevich, Aleksandr Aleksandrovich Teoreticheskiye oanovy radioavyazi (Theoretical Bases of Radio Communication) Moscow, Gostekhizdat, 1957. 7~41 p., 25,000 copies printed. Ed.: Kostiyenko, A.I.; Teeh. Ed.: Oavrilov, S.S. PURPOSE: This monograph is addressed to third-year students in radio communication engineering institutes. COVERAGE: The material presented in the book constitutes a course designed to provide the students with the fundamentals of radio theory and technique as applied_&o all. Ru eqxWjat disciplines in radio engineering, such as~-"dbiveii, aimp.lifiers,-tranmitt-drs, radio broadcasting, television, etc. Practical problems of radio engineering are not treated in the present work. The circuit diagrams, and certain structural presentations,,therefore, are given-only as illustrations of general principles. The material presented in this book differs from the preceding course on the Card 1/7 Theoretical Bases of Radio Communication 476 fundamentals of radio engineering in that it contains a special section on the fundamentals of radio communication. Some of the classical material (the section on the theory of circuits, for example) has been eliminated from the present work to make room for new material. The course of study presented in this monograph corresponds on the whole, to the program adopted at the 1953 Conference of Representatives of the Moscow, Leningrad and Odessa Institutes and later confirmed by the Administration of Schools of the Ministry of Communications in August,1955. This course of study consists of three parts, the first two of which are embodied in the present work. The third part of -the course was published in 1956 as a separate book by the same author. 'There are no references. TABLE OF CONTENTS: Foreword PART I. FUNDAMENTALS OF RADIO COMMUNICATION 5 Card 2/7 Theoretical Bases of Radio Communication 476 Ch. 1. Communication and the Signal 7 1. Communication: perculiarities of radio communication 7 2. Block diagram of-a commiinication'cir6ilit 9 3. Concept of radiation and the propagation of radio waves 10 4. Physical characteristics of the signal 14 Ch. 2. Modulation and Codes 21 5. Modulation: general concepts 21 6. Spectra of modulated oscillations 28 7. Vector diagrams of modulation 37 8. Pulse modulation 39 9. Code: general concepts 43 10. Kotellnikov's theorem 4T 11. Quantization 52 12. Pulse-code modulation 56 13. Multichannel communication 58 Card 3/7 Theoretical Bases of Radio Communication 476 14. Frequency and time division 60 Ch. 3. Noise and Noiseproof Feature 67 15. Noise: general description 67 16. Fluctuation noise 69 17. Probability description of noise and signal 75 18. Noiseproof feature 79 19. Methods of preventing noise 83 20. Relative noise ImmunitY of amplitude and frequency modulation 87 PART II. LINEAR FHEN014ENA IN RADIO ENGINEERING Ch. 1. Problems and Methods of Linear Theory 94 21. Linear phenomena systems 94 22. The superposition principle 96 Card 4/7 Theoretical Bases of Radio Communication 476 54. The round waveguide 325 55. Excitation and filtering of waves in the wave- guide 330 56. Losses in the waveguide 336 57. Cavity resonators 342 AVAILABLE: Library of Congress JJPAsv 8-4-58 Card 7/7 AUTHOR-: Kharkevich, A.A. 264 TITLE: Possibility of signal spectrum compression. (0 vozmozh- nostyakh szhatiya spektra signala). PERIODICAL: "Elektrosvyaz"' (Telecommunications), 1957, No.4, April, pp-3-11 (U.S.S.R.) ABSTRACT: The purpose of the present article is to clarify at least a few of the basic assumptions underlying the problem of signal frequency spectrum compression. This problem has lamely been given much attention but most works bear more onto the practical rather than the theoretical side of it. The author begins with general considerations governing the input and output signals, the amount of information and with the theorem of conservation of information. The non-statistical possibility of spectrum compression is next considered and two examples of spectrum compression of the signal with information: I = 2FT log m are given, where F is the spectrum bandwidth, T is the duration of the signal, &t - the interval between &t symbols, expressed by the Kotelnikov theorem as 1 and m is the equiprobable value of the quantised signal x(t). The author then proceeds to consider theoretically the possibility of spectrum compression by changing the signal statistics. He establishes that F Possibility of signal spectrum compression. (Cont.) 264 may be compressed, for a given constant T, by increasing the entropy I,; this would lead to the elimination of internal correlationst which in turn, would equalize the probability of independence of separate symbols. Since the internal statistical correlation decreases the entropy and since then, for transmission of a given information, a signal with a large FT product is needed, this leads to redundancy R. Two formulae for R are considered: It Tr_ 0 where V = log m and 0 FOTO where F and T are actual spectrum bandwidth and duration of the signal and F and T - are same but for the signal with the s9me inforgation but for maximum value of V = V. The redundancy of AM and FM Signals is next considered. The, redundancy R for AM is being shown to be 0.5 and for FM to be 0,97 (in both cases they are eliminated eventually by detection. The author also mentions that the reduction of noise inteff6rence in FM signals, due to the increase in /7 - /7' INFORMATION THEORY "Cri the Theoretical Cptimuti Corwinication Syste-,:i" by P.A, MarIkevich, Elektrosvya_Z', Vo 5, MAY 1957) PP 15-18- From the theoretical point of view, the choicc of a cornizAcation systcm reduces to the choice of a mathod of transmission (i - e., ccle), and of a method of reception. An optitium sySten s:,,ould give the be3t reproducibility for a specified noise rejection, or conversely, the best noise refection for a specified reproducft:~lity. It- is shovn that obtaining the optinun, syctc-m reduces to a certain variational problerm, the fornulation of w11,1ch eciiinains the noise-dis- tribution probability. The itatement of the problem is illustrated with ~-cveral examples. Card 1/1 MON I 8 - 29A CARD 1/1 PG 832 SUBJSCT USSR/Mk.L ICS/Geometry AUTHOR CHARUVI A.A. TITLE A problem. PERIODICAL Uspechi mat.Nauk 12, 2, 184 (1957) reviewed 6/195T The author gives the following problem: Determine the number of corners of a more-dimensional cube the distances of which from a given corner and one to another are not smaller than a given magnitude. AUTY.CiR. 1harkevichi A.A. .L Ij S arlson of Several Possibil"'tiec Se:adinC Simple TIT I;B Pi-,tu'es (S ic,zhnostey p,~,redachi - r ravneni ye nekotorykh vozm prostykh risunkov) PERIODICAL. Blektros-ryaz' , 10/58, Nr 5, PP 44 - 4? (TJSSR) APSTRIM! The previous work of Shannon, Loeb and Benjani'M is 1,qent-Loned" Conventional rit,ethods of sendinf, si.:ftple line d_ra,.,.'d_a,-,s are ineffic.ient; sin~;e the area r-owred by the of the d-rawing is a minute fraction of the total area of the figure. T-wc, suggestions are examined, in one, the co-ordinates of several noints in the course of the lines are transmitted; in the other, the scanninU is --arried out alone,-,- the factual lines therasel-,-es~ The improvement factor ~ratio o -v Voluz-_es in -_;onventicnal and -pro-Dosed wethods) for each case is Civen in and (2), respective3y, In oracr to compe-re the t- m, o sy,,;tem-,t , it i s aFsuwed , ii)t I I it iVCly 7 UMt U,C 11U,11:110r Of 0 ,'1 MbL)1E3 trail-smitted is linenrly related to the total lentth of 1.inQ in the drawing. The se.,ond system is r:.~Dre ef'L'c-~~t,_ve than 1.1~e f.,-rst by a fc,,:.tor approximately equai to '~ Lc L,11) /3, TIh e f i r s t-- w;--thod makes inadequate use of the statisti_*~s of tte zessage ,)ul; the :3econd encounters particular diffi2ulties in Card 1/2 r, C:f SP-'fe-ral Possib.-ilijties ir 5E- p 'rhe drawing does be so P no+ consist I cf a conti,-Iu,)us 1-ne, it r eq .ust -repared 01' scrutinise-d With a rather r, ulDmont including information GtOraze ca-a,~i4-y -.-.her hand, q - the fixEt method c On ~,he repeated transmission with an be ~,ade `,'re effecrive by r 'Jif f e by f a rent dire",'4_iOn cf scan. or ral irst storing the information and then rejuC,'AS-r1_[..' i It at a which wakes better use of the channel Capacity. The se,-cnd method may also be improved by 7akin,- 'L Cder Probabilities, use Of h her 0 for exaMple, that the nLumbcr andL-';.-ind of rianr-es in dire,-'tion that the'lii'le may make is in many cases '--r,7 -restricted. There are 3 figu-res and 2 references V.-i--h. SOViet and 1 EnGilsh. of 8UB1',',ITTED- Uard 2/2 , February 24, AUTHOR: Khar 42XL9Z,. A - A. SOV/106-58-11-1/12 TITLE The kossibilities of Spectrum Compression (0 vozmozh- nostyakh szhatiya spektra) PERIODICAL: Elektrosvyazl, 1956, Nr.11, pp.3-8 (USSR) ABSTRACT: At the present time the problem of spectrum compression has still not been completely solved due mainly to theoretical difficulties. There is however one class of process about which one can say something rather definite and which may be useful for the further development of theory and technology. This is the class of modulation processeao In human speech sub-carriers are formed both by,the expulsion of air from the throat and by the vibration of the vocal cords. These carriers then support quite narrow-band modulation spectra. In the system of trans- mission considered the original spectrum is analysed by ~'bank of narrow-band filters whose outputs change slowly with time. Eq.(4) gives the instantaneous spectrum from a typical filter. At the receiving end, for reconstituting Card 1/2 the original process, a noise generator feeds a bank of The Possibilities of Spectrum Compression. SOV/106-5a-ll-1/12 filters and a system of amplitude modulators which control the strength of the outputs from each filtero A further simplification is possible if we replace the noise-fed narrow-band filter by a simple sine-wave oscillator. The problem now reducas to the transmission of amplitude and instantaneous-phase modulating functions. it is concluded that the use of simple sinusoidal signals for reconstituting speech is quite inadequatep ai-1 this is borne out by the experiments of Marku and Dage (Rof.6). There are 6 references, of which 4 are Soviet, 1 English and 1 German. SUBMITTED: June 30, 1958. Card 2/2 AUTHOR: Kharkevi 103-13-8-1/12 TITLE: On the KotellniYov Theorem (0 teoreme Kotellnikova) survey of Some of the Latest Papers (Obzor nekotorykh noveyshikh rabot) PERIODICAL: Radiotekhnika, 1958, Vol- 13, Nr 0, pp. 3 - 10 (USSR) A13STRACT: The Kotellnikov theorem (abroad it is called the sampling theorem) dating from 1933 is the basis of all pulse signaling systems. An accurate relation it is only valid for functions the spectral density of which is equal to zero for all frequen- cies outside a certain finite intervalg the spectrnm width. The transmission functiont however, represents a random process; a random process the spectral density of which is equal to zero in the finite interval is a singular process, i.e. the process values can be predicted with any accuracy for any tine. This again means that'processes with strictly limited spectrum can not transmit informations, i.e. that no process representing a signal can have a finite spectrum.- For this reason some papers were published in 1956-1957 which dealt with Kotelinikov's theorem. It would be the most natural thinC to reZard the Card 1/3 theorem not an an exact statement referring to a function with On the Kotellnikov Tneorem. Survey of Some of the :~',,V/1013-13-0-1/12 LrAest Papers finite spectrum but as an approximate assertion referring to a function with an infinite spectrum. This opinion in uttered in the paper by I.T.Turbovich (Ref 3). The evaluations of the error by the approximation (Refs 3 and 4) rLaree with each other. In the paper by Zheleznov (Ref 5) a more generalized problem in posed from the very bo,,,,innin(;: the beat representation of the randora function u(t) I)y an expansion accordlnLm- to the method used in formula (10) is to be found. In the ALA-scheme a physi- cally nimilar expansion is reali-.,od. It reals: continuous quasiateady signale u(t) with an infinite spectrum can be trans- mitted by means of fi,-ures followinE; one another everyU 1- se- condo with an accuracy arbitrarily close to the final accurracy 90 , when the periodIr 1 does not exceed the correlation inter- val'r 0 and the siCnal duration T is much longer than the correlation interval IC 0. This formulation comprises the recults of all papers. Only the theoretical inforTation content of the Kotellnikov theorem remLined uninvestiCated. This -ap was closed Card 2/3 by the lecture delivered by KolmoCorov (Ref 6). U On the Kotellnikov Theorem. Survey of Some of the 108-13-0-1/12 Latent Papers Concluding the author states that the Kotellnikov theorem is at present regarded as an approximate (and not accurate) assertion, and that it makes possible the following statement: 1) An approximation formula for the process in the form of a series of lagging functions, and 2) to give an approxinate evaluation of the number of measurements of the random vector proceeding from the desired accuraoy in the determination of the E -entropy. There are 1 figure. and 6 references, which are Soviet. SUBMITTED: February 24, 1958 1. Radio signals--Theory 2. Scientific reports Card 3113 AUTHOR: Kharkevich, A. A. SOV/108-13-9-14/26 TITLE: Letter to the Editor (Fislmo v redalctsiyu) -PERIODICAL: Radiotekhnika, 1958, Vol. 13, 11r 9, pp, 69 - 69 (USSR) ABSTRACT: This letter su--ests the compilation of an atlas of typical pulse generating circuits. First the ordinary block diW&M if various complicated devices are listed. Then the principal circuit diagrams are compiled. This is where this atlas would be of great help. This atlas should contain a complete presentation of the most modern circuit diagrams of individual blocks, which have been examined and which already have stood their test. Only a minimum number if variants should be incorporated, giving technical information and data on well established modes of operation. The next stride in this direction would be the industrial production of types of circuit elements. Elements for such a rationalization are already in existence in the USSR and abroad. Card 1/1 ich, red.; RAYEVSKIY, S.Ya., red.; ov I ~ch '"~AKH=L.AAHOOVV. ~SG.11 .0 . - ~tteekkhhnnq ~rree 0v.0 Grif ormation theory and it's applications; collection of translations from the English) Teoriia informataii i se pri lozheniia; sbornik perevodov. Pod red. Khorkevicha. Rosk-~;~, Gos.izd-vo fisiko-matem.lit-ry, 1959. 328 p. (KIRA 12:12) (Information theory) AUTHOR: Kharkevic2:.., sov/lo6-59-5-11/13 TITLE. e er to the Editor: Photo-Telegraphy from the Point of View of Telegraphy (Fototelegraf a tochki zreni.ya telegrafa) PERIODICAL: Elektroavyazl, 1959, Nr 5s pp 73-76 (USSR) ABSTRACT: The object of this letter is to compare the potentialities of black-white photo-telegraphywith normal telegraphy for transmitting written information. It is shown that telegraphy is more efficient than photo-telegraphy. For the purposes of comparison, the original information is assumed to be in the form of type letters situated in rectangles of 1.5 x 2 nim dimensions. if the resolution of the photo-telegraphic apparatus is such that the finest detail is a 0.2 wm square, then the number of elementary squares per letter is 75, requiring 75 binary digits (current pulse (1) for black; no-pulse (0) for white) per letter. However, it cannot be concluded that the Baudot code which use3 5 digits per letter is therefore 15 times more efficient than the photo- Card 1/5 telegraphic method, since photo-telegraphy has greater sovio6-59-5-11/13 Letter to the Editor: Photo-Telegraphy from the Point of View of Telegraphy interference-stability. For true comparison, it in necessary to reduce both systems to a common condition of interference-stability. The author considers the construction of letters from a grid containing 3 x 4 =.12 elementary squares. LYhis number is not sufficient for all the letters of the Russian alphabet but this is not important for the author's argument.) Letters with the simplest forms, constructed in this manner, are shown in Fig 1. Due to interference, errors arise in which a no-pulse digit is replaced by a pulse digit (or vice versa). This is called a single error. If two digits are incorrect, it is called a double error and so on. The smaller the probability of replacement of one letter by another, the greater the interference- stability of the system and, in its turn, this probability is smaller. the more the letters differ from one another. This probability can therefore be qualitatively expressed by the number of different Card 2/5 digits in the different letters, The number of SOV/106-59-5-11/13 Letter to the Editor: Photo--Telegraphy from the Point of View of Telegraphy different digits is called the spacing. The spacing between the different letters is not constant, the spacing between the seven simplest letters considered being shown in Table 1. Letters having the greatest spacing with respect to tile otheys, e.g. the letter T, are the most atable. However, the spacing alone is not. sufficient; two different letters can differ from a third by the same spacing, the difference lying in the "direction" of the elementary squares, i.e. on the direction of the scan. If the bla~:.k elements are denoted by I and the white by 0 and the scan of the letter rectangle is as shown in Fig 4, then a 12 digit binary codeas shown in Table 2, is obtained, The spacing increases with the number of digits. thereby increasing the interferonce-stability but dtcreasing the efficiency of the transmission. It is possible to increase the spacing by departing from the normal letters and using new simplified configurations satisfying the Card 3/5 demand for maximum spacing but this, in fact, is the sov/1o6_59-5-11/13 Letter to the Editor: Photo-Telegraphy from the Po_4nt of View of Telegraphy principle of the telegraph system. The usual 5 digit Baudot code is next considered. The inininjum spacing between two code combinations is unity. ThuR, for a single error, a different letter is received and the error remains undetected. A simple error detecting code can be obtained by adding one more digit and making the number of O's or I's in any code combination an even number. The least spacing in then two and this code with 6 digits ix as effective as the 12 digit photo-telegraphic code considered earlier. Using the same scan sequence an shown in vig 4, "images" of the letters can be constructed. The resulting images are shown in Fig 5 and the letter spacings are tabulated in Table 5- Comparison of Tables 1 and 5 and of the images permits direct comparison of the telegraph and photo- telegraph systems. Table 5 shows the spacing of the telegraph system to be equal to, or relatively better than, the spacing of the photo--telegraphic system Card 4/5 (which refers to a 12 digit code). It is concluded that sov/lo6.,59-5-11/13 Letter to the Editor: Photo-Telegraphy from the Point of View of Telegraphy in telegraphy an optimum code can be chosen independently, whereas in photo.-telegraphy simplification of the letter forms leads to longer code combinations for the same interference stability and hence to less efficiency compared to 'the telegraphic system. SUBMITTED: 19th January 1959 Card 5/5 -rig 24846 8/106/60/000/004/001/007 6:2r-00 A055/A133 AUTHORS: Blokh, E. L., and Kharkevich A. A. TMU: Antifading coding PERILMICALi Elektrosvyazl, no. 4, 1960, 3 - 6 I.M.: A method of signal transmission is described, uzing correcting anti- int,erference codes and allowing to enhance the reliability of communications in the presence of fading. Assuming that the transmitted communication is coded by n-digit combinations of a uniform code, a group of N such combinations is taken and written down as shown in Table 1, number N being chosen so that the time of transmission of N binary digits should be sufficiently long compared to the aver- age duration of fading. Transmitting Table 1, not by columns, but by horizontal. lines, a part. of the transmitted signal will vanish owing to fading. Replacing the vanished digits by an asterisk, we obtain Table 2 for the received signal. If -the received digits are now grouped according to columns, we obtain code com- binations from which certain individual digits have vanished. If N - and this is the essential point - was chosen in conformity with the statistics of fading, the disappearance of an individual digit from the combination can be considered as Card 1/~ Antifading coding 24846 s/106/60/000/004/001/C/J7 A055/A133 an independent event. The digits that vanished owing to fading are distributed in a random manner in code combinations. If each column contains one combinatiOn: the errors in the same positions in the adjacent combinations are stro-ngiy corrg- lated. L:Ut if, several combinations, representing a certain section of the com- IranIcation, are placed in one Column, the error can already be, considered as in- deperident, not only within the given combination, but also within the limits Of the corrimani cation section. In the case of an additive interference, a certain In the case d-igit is replaced by an erroneous one (e.9. 0 by I or vice versa). of a multiplicative interference of the fading type, the digit is not replaced$ I MU ;t-red, it, Lhan r vanished digits st be re, .,ut vanish-is altogether. If not more is sufficient to use the code with a distance between combinations at, leastequal t,7, r + 1. Comparing the received combination with all possible ones, it Can De Feen teat the received combination coincides with the transmi"'ted one and differs all other cambinatio~ns in at least one digit. Trne transmitted -ombination ran t,hu3 be identJfied and, consiqueutly, all the va-nished diwits can be restozed. -:f tne same code is used in the pre9ence Of an additive interteritnee, it- will Tnc-rsly allow to detect errors whoze number does not exceed r '-, it will not allow be impossible. -T-ne interference- thern, and their correction will thus kiilir4 feature is characterized by the probability -,f an errol-fl,c-e recepti-in )f Car,l -2/5., ~Y Ant,ifading coding 24PI16 5/106/60/000/004/001/007 A055/A13_3) a sequence of L elements of the communication. The ratio-9 between the duration of the vanishing of the signal and the total transmission duration can serve as the parameter determining the fading action. When no correcting code is used, the probability of error-free reception of a sequence of M digits is P, 0 0 or, for I ~M p, e When a code restoring not more than r digits in-each n-digit code-combination is used, the probability of correct reception of each combination is I _ Cr_P1 r+1(1 - ~)n-r-l n n E and the probability of error-free reception of the section of M digits of the initial sequence will be n M r [l Ck ;k(l -)n-k]m P2 k-r+l n or, for E 1 M Cr+1 r+I P(r) e ill n 2 Card.3/~!/ (2) 248h6 S/106/60/000/004/001/007 Antifading coding A055/A133 In particular: (R-1)" +1 P~') 2 2 (3) p.(12)e 3! all. (4)' A comparison of (3) and (1) shows that the use of the code that restores one.di-' git is expedient if: M+l 2 A comparison of (3) and (4) shows that the code restoring two digits must be us';a only if: n(n-1)(n-2) 3m(m+l) There are 2 tables and 5 references: 4 So-&t-bloc and 1 non-Soviet-bloc. The rer- erence to the English-language publication reads as follows: Price, Gr6en. ~.Acom- munication technique for multipath channels!'. Proc.' ME., 46, no. 3, 1958. f3MUTTED: November 24.. 1959 Card 4/5(1 27633 S/19X61/000/002/013/039 *Awl 76 D216 302 iWTHOR: hharkevich, A.A. TITLE: The value of information PERIODICAL: Referativnyy zhurnal. Avtomatika i radioelektranika, no.2, 1961, 33, abstract 2 V252 (V sb. Probl. kiber- netilci, no. 4, M., Fizmatgiz, 1960, 53-57) TEXT: Information is usually collected for a certain spccific purpose and it seems therefore logical to raise the question of ~he value of information as depending on how this information helps in obtaining this purpose. The same information may have different values, depending on the aim. Existing theory ignores the meaning, of information and its value for the addressee. In the case when the aim - for the attaining of which the information is being collected - can be clearly determined, the valiie of the infornia- tion can be determined as the increase in the probability of attai- ning the aim. If, up to the instant of receiving information, this Card 1/2 The value of information 94 61 000/002/013/039 S/I D216 30 probability is po and after the arrival of information this proba- bility became pl, then the value of the received information c;ai be determined as Pl 1092 PI - 1092 Po = 1092 Po It is shown in an example that the value of negative when it decreases the probability It is suggested that it be called, in this A few examples of evaluating the value of 2 references. information can be of attaining the aim. case, disinformation., information are given. Card 2/2 N . . ly .. KH;xRMICH, A.A. In regard to professor M.S. Heiman's article. Izv. vys. ucheb. zav.; radiotekh. 3 'no.4'519-520 Jl-Ag 160. (MIR& 13:10) 1. MoBkovskiy elektrotakhnicheakiy institut evyazi. (Radio--Study and teaching) S OV /I 1/ 12 AUTHOR: Kharlc~Vich, A. A. TITLE: &n Principles of Readir..r Machine ~T- PERIODICAL: Radiotekhnillca, i-960, Vol 15, pp -A-9 (USSR) ABSTRACT: The paper discusses some geneval problems of construc- tion of reading machines. The definition of a reading machine Is given as a machinc which -~u,6omatlcally recognizes letters, or othe.- o!' a pi-Inted or, written text. Any reading machluo ha:3 to perform the follo',-iing basic operations: (a) presen-,ation and examination; (b) preparation of description; and (c) comparison of the description .,itth a standard, i.e., the proper vecognitiori operation. The examination consists In the action of the ureserit-ed oattern on a transducer, normally of the photoelectric type. There- by a corresponding electric sic-rjai Is prrod-ucedl. The desc-10tion Is the most lmnartlan, o-.c-atlon. I t cons I sS In f orm trig a s Ignal deoc rInil-1- the pa!'- tern Card 1/6 4 In a rrianner, most convenient foz., vecognition purposes. Or, Pi-Iric ltle_- of Rea,l 1-.;7 Mac- Ine 1,,onstruc %lot-,, s 0", 1 1 card 21c', Opei-at-lon.- (a) and (L)) ma, ue (,,,eratlon In princ.1ple, docs noel diffleultles. It Is obvlouo --0a1_'_ln,, mac'nlr%e Mku"t, have some kind of' m(nno-,- !,_i .he :,I p t I or u3 a,, k3 In a 11,~rco OC veu~iinig muchInes the examination Is r-,iaille hy scannIng, 'ahereby the readIng follows a trajectoi,~, ~ In the plane of the pattern . This trajectory is 0- composed of' separate sections. 'rn, e ~;2annjng 31q;nal may also be used for the deocrip,_'o~:. -It is Iniportant in thiz case that an optimal scannin- selQctlon be made. Two kinds of deocription ai,e consIdeved: (1) absolute descriptIon, pevirilttinf,, ~(,,jluorat~on of the de3~;rlbed object with ceptain accut,ac~, ; (2) reiative description, containing only the features by which a certain object differs from other objects-of the same set. The latter description Is sufftcLent for recog- riltion purposes and is more economl-cal. An optllrriuum descrIptlon should be 'he slorteo' ol~-,,ce a z;liorter descriptlon meano a T 1,. C autlhov j-,Ives examples Of it '77 - r-, ~ ' On PrInc1ples of Reading Cori ~3 1, rl-ic t Ion I i , U S011110;3 - 15 -2 - 1 /12 lie points out, however, that the probicir, of minliTilzation of absolute as well as of relative descriptions is not only still unsolved, but has no"'l- even been suffIciently clearly formulated. The simplest scannin'c', methqds are in direct relation with a certain system of coordinates. The usual TV scanning, for example, is related to the rectangular system. There is a special type of scanning called follofi-up scanning. lieve, the scanning ray follows the contour of the pattern, and thecut,ve itself is used as coordinate axis. One of the de- scription methods consists in the use of the scanning signal for description. There is another method, called topological descrilption. This method consists in counting the topological features of the pattern. Topological description has a rather general Qhavacter. However, it Is not sufficient for the recognjtiQn of letters and digits which are topologically identical-. e.g., 6 and 9. Therefore, some elements of geometrical description must be added. Since letters and digits of various forms and sizes must be recognized, the problen Card 3/6 On Pri.nciple"! of' Card 4 /6 trar,Lt'oruiatlons 13 of' great P, 11-~ ct-ated that !,r.(-)In -a prai.-tical point ot, v"Y"'. "'011C."l-ur, scannLrlt-, to an lmpoi~tant de_,cv1p1,-1on It does no,t vary with respect to the actual!~., uc~!,._Jrring trans- fVvmations. Topological descriptlo,; also ~ioes ncLt, vavy . I lowe v c i-, I t,-, t e c hn I q ue I ~, :, t I L I c omp I e x .The recognition may be an operation slmuLtanco~is or stepped- up. When t'he pattevris to be rccoj~nLzcd are character- ized by n features of which an~ ilia-, be either present or absent, then the description of a pattern is an n-digrlt binazy number. In case Of' SIMLIltaneQus recog- nitLon, n-digit numbers are compare,1. In case o' a stepped-up recognt~ion each of the binary' numbei' Is compared step-by-step, thus okeratLng with (ine-digit numbers. This leads to simplification of -uhe machine. The stepped-up method permits descriptions of varying length. However, simultaneous description is more nolseproof than stepped-up description. The recognitiOn Is the more reliable, the more the dlesc rip t ions of thcf pat-terns differ from one another. The (A'C33CriPtI0r1 JIC- ferences may be associated with the concept of distances. On Principles of Reading Machine Construction 77556 t(Y11/108-15-2-1/12 Card 5/6 Thus, in addition to the above requirement of minimization, a description must satisfy the require- ment of a greatest minimum distance between two de- scriptions. It is not clear If this distance maximization may be formulated as a mathematical problem. The actual printed characters need some preparation, which consists in elimination of minor faults. The preparation may be a separate operation or may be combined with the examination. On the subject of the so-called self-teaching reading machine, It Is stated that this prepares the standard description by Itself. This Is done under control of a human operator, from which t'he machine receives signal-s of approval or disapproval. Under the action of these signals the description is corrected. When a sufficient degree of reliability is achieved, the self-teaching process is terminated. The paper concludes that a detailed classification of reading machines must be postponed until unsolved problems are cleared up. By finding optimum solutions, the still very complex technique of reading machines may be simplified considaably. 63149 S/108/60/015/009/001/OCB B002/BO67 AUTHORSt Blokh, E. L., Kharkevich, A. A., Members of the Society I TITLE: Some Properties of Communication Systems With Fading PERIODICALi Radiotekhnika, 1960, Vol. 15, No- 9, PP. 3-9 "Y TEXTi Only additive fluctuation noises have as yet been theoretically studied. The signal received is regarded as the sum of the emitted signal and the noise. On the other hand, multiplicative noise, the so-called fading, has hitherto not been theoretically treated. It consists in the fact that the intensity of the received signal is subject to random fluctuations. Three cases of interference are distinguished in the theoretical treatmentz 1. Additive noise. II. Fading. System with an active pause (phase modulation or frequency modulation). III. Fading. System with a passive pause (amplitude modulation). The dependence of the carrying capacity on the transition probability is computed for each case and graphically represented (Fig.). Furthermore, the identifica- tion of the regenerative codes is considered. As computation shows, case I requires distances as large as possible and code combination as long as Card 1/2 83909 S/10 60/015/010/002/008 B01 2YB060 AUTHOR% Kharkevich, A Active Member of the Society TITLE: The Discrimination of Continuous Signals~3 PERIODICALt Radiotekhnika, 1960, Vol- 15, No- 10, pp- 11-13 TEXT: The author examined the possibility of representing by one or several numbers a signal given in the finite region in the form of a con- tinuous function., The optimum selection of such numbers is discussed. The present considerations were made in connection with the construction of identification machines serving for the identification of continuous circuits. If in the range of 0