CA1276688C - Method of digital signal transmission - Google Patents

Abstract

ABSTRACT OF THE DISCLOSURE
A modulation method capable of high-speed or high-quality digital data transmission under a multipath fading transmissional, such as a radio transmission in an urban area of differential coding phase modulation and resides in the fact that the rate of change of phase varies, or the phase is discrete, but does not employ a phase transition waveform of a constant value in a time slot which is a smallest unit for transmitting one symbol of data as in conventional methods. Phase transition waveforms in respective time slots which are spaced apart from each other by prescribed time slots are identical varying waveforms or discrete waveforms, and entirely shifted from each other by an amount according to the data transmitted. A detecting method employed is a differential detection method using a delay line for delaying a signal for a time corresponding to the prescribed time slots. In the presence of multipath propagation, a plurality of kinds of detected outputs are produced according to the multipath propagation, and the detected outputs are combined by a filter after the differential detection, producing a diversity effect for improving bit error rate characteristics. This digital signal transmission method is capable of multiphase transmission and can increase spectrum utilization efficiency.

Description

lZ'~6688 TITLE OF THE IN~ENTION
METHOD OF DIGITAL SIGNAL TRANSMISSION

BACKGROUND OF THE INVENTION
1. Field of the invention:
The pr~sent invention relate~ to a method of transmitting a digital signal along a multipath fading transmission line such as a radio transmission in an ur~an area.

2. Description of the Prior Art:
In recent years, digital signal processing techniques have spread to the field of mobile communications so as to meet increasing demands for communication privacy, intelligent communication, a~d matching with peripheral communication networks. In urban areas where such demands are most intensive, the quality of communication is considerably lowered by multipath propagati~n due to reflection and diffraction caused by buildings and surrounding geographical features. In digital transmission, when the propo~ation delay time difference between waves constituting the multipath is increased to such an extent that it is no longer negligible with respect to the length of a time slot, the bit error rate characteristics are highly degraded by waveform distortions and follow-up failures in synchronous systems.
FIG. 1 is a phase transition waveform diagram showing an example of phase transition waveform of a transmission signal transmitted by a digital signal transmission method according to the present invention;
FIGS. 2 through 12 are phase transition waveform dia-grams showing examples of phase transition waveforms of trans-mission signals transmitted by a digital signal transmission method according to the present invention;
FIG. 13 is a phase transition waveform diagram showing a specific example of phase transition waveform of a transmission signal transmit~ed by a digital signal transmission method according to the present invention;
FIGS. 14 through 18 are phase transition waveform lZ7~i688 diagrams of other examples of phase transition waveforms in a time slot of FIG. 1;
FIG. 19 is a block diagram of a generator circuit for a transmission signal according to a first embodiment of the present invention;
FIG. 20 is a block diagram of a circuit arrangement of a quadrature modulator 1905 as shown in FIG.
19;
FIG. 21 is a block diagram of a circuit arrangement of a differential coding circuit 1902 as shown in FIG. 19;
FIG~ 22 is a block diagram of a circuit arrangement of a wveform generator circuit 1904 as shown in FIG. 19;
FIG. 23, is a block diagram of another circuit arrangement of the waveform generator circuit 1904 as shown in FIG. 19;
FIG. 24 is a block diagram of a differential detector of a binary-phase system;
FIG. 25 is a block diagram of a differential detector of a quadrature-phase system;
FIG. 26 is a block diagram of a differential detector of an octal-phase system;
FIG. 27 is a diagram explaining a detected output signal in the presence of a two-wave multipath propagation in a~digital signal transmission method of the present invention;
FIG. 28 is a vector diagram illustrating the combined phase of a D- and U-waves in order to determine the detected output in FIG. 27;
FIG. 29 is a block diagram of a diversity model in the presence of multipath propagation in a digital signal ~ transmission method of the present invention;

12'7~6~8 FIG . 3 0 is a diagram explanating a detected output signal in the presence o~ a two-wave multipath propagation in a digital signal transmission method of the present invention;
FIGS. 31 through 35 are graphs showing average bit error rate characteristics according to the present invention in the presence of two-wave Rayleigh fading FIG . 36 is a block diagram of a transmitter circuit according to a second embodiment of the present invention;
FIG. 37 is a phase transition waveform diagram showing the phase transition of a transmission signal transmitted by a first conventional digital signal transmission method;
FIG. 38 is a diagram explaining a detected output signal in the presence of a two-wave multipath propagation in the first conventional digital signal transmission method;
FIG. 39 is a vector diagram illustrating the combined phase of a D- and U-waves in order to determine the detected output in FIG. 38;
FIG. 40 is a phase transition waveform diagram showing the phase transition of a transmission signal transmitted by a second conventional digital signal transmission method;
~ - FIG. 41 is a block diagram of a detector in the second conventional digital signal transmission method;
FIG. 42 is a diagram explaining a detected output signal in the presence of a two-wave multipath in the second conventional digital signal transmission method; and FIG. 43 is a vector diagram illustrating the combined phase of a D- and U-waves in order to determine the detected output in FIG. 42.

668~3 A first conventional digital signal transmiscion method will be described, by way of example, with reference to FIGS. 37 through 39~

FIG. 37 is a phase transition waveform diagram showing the phase transition of a signal transmitted by the first conventional digital transmission method. Designated at T is a time slot interval which is a minimum unit for transmitting one data symbol. When the data is a 1, a 180 phase transition occurs, and when the data is a 0, no phase transition takes place. This modulation technique is called differential coding BPSK (Binary Phase Shift Key~ng).

Such a transmitted signal can be detected, for example, by differential detection (delay line detection) using a delay line with a delay of one time slot. As a typical example of multipath propagation, consideration will be given as to how a detected output signal behaves in the presence of a two-wave multipath transmission line having a propagation delay time difference which is not negligible as compared with the time slot interval. A wave arriving earlier will be called a D~wave, and a delayed wave will be called a U-wave.

FIG. 38 is a diagram explaining how the detected output signal behaves when the transmitted signal as shown in FIG. 37 is subjected to the differential detection in the presence of the two-wave multipath propagation. FIG. 38(a) shows a phase transition of the D-wave. A phase transition of the U-wave which arrives with a propagation delay time difference from the D-wave is illustrated in FIG. 38(b).
The detected output at a particular time is the vector inner ~ product of the combined phase of the two waves at that time ~Z~66~3 '- . ..

waves at that time and the combined phase of the two waves in a preceding time slot. For e~ample, the detected output in a region B in FIG 38(c) is the vector inner product of the combined phase of the two waves at a time B' and that at a time B.
FIG. 39 shows the combined phase of the D- and U-waves in order to determine the detected output at each of the times A through C. The ratio of the amplitudes of the D- and U-waves is indicated by p, and the phase difference between them by ~. For example, the absolute value of the detected output at the time B is the inner ¦ product of vectors OB' and OB in FIG. 39, i.e., the square of the line segment OB. Therefore, the detected outputs at ¦ the respective times A through C in ~IG. 38~c) are given, using the cosine theorem, as follows:
A ... indefinite B ... an~l + p2 + 2p cos ~) C ... indefinite i whe~e an (an = + 1) is a data sequence being transmitted.
¦ In the regions A and C~ the detected outputs are indefinitedepending on the data valuec in the preceding and subsequent time slots. After the differential detection, a low-pass filter is normally placed in order to remove 1 harmonics and undesired noise components Therefore, the I waveform o a final detected output signal is as shown by I the dotted line in FIG. 38(c), which is produced by filtering the solid-line waveform of ~IG. 38(c), and lZ76688 constitutes a portion of an e~e p~ttern. If p is close to 1 and a is about 180, the detected output in the region B
which is an effective detected c~ltput is substantially zero.
Thus, the eye is closed, and the bit error rate characteristics are degraded At this time, since the ineffective detected outputs in the regions A and C are much larger than the effective detected output in the region B, the eye is largely fluctuated in the direction of the time base, making a reproducing clock unable to follow it and thereby resulting in a greater degration of the bit error rate (See, for example, "Bit Error Rate Characteristics in Rayleigh Fading Having a Propagation Delay Time Difference"
by Onoe et al, Papers of Tech. Group on Commun. Syst., IECE
JAPAN, CS81-168, 1982, or "Analysis of Instantaneous Bit Errors due to Multiwave Propagation and An Error Generating Mechanism based on A Bit Synchronous System" by Takai et al, Papers of Tech. Group on Commun. Syst., IECE JAPAN, CS83-158, 1984).

In order to reduce the degradation of the eyes pattern and the fluctuation of the eye causing the degradation of the bit error rate characteristics, there has been proposed a method such that the phase transition waveform of a transmitted signal is designed so as to produce a plurality of kinds of detected outputs and the problems are improved by a diversity effect produced by combining the plurality of kinds of detected outputs. One example of such a second conventional digital signal transmission method will be described below with reference to FIGS. 40 through 43.

~276688 FIG. 40 shows the phase transition of a signal transmitted by the second conventional digital signal transmission method. T indicates one time slot for data.
When the data is a 1, the phase is rotated twice in one d~rection, each by 90 in every half time slot. When the data is a 0, the phase is rotated twice in a different direction from the above, each by 90 in every half time slot. This modulation technique is called DSK (Double Shift Keying).
Such a transmitted signal can be detected by a differential detector having a delay line with a delay of half time slot as shown in FIG. 41. Denoted in FIG. 41 at 4101 is an input terminal, 4102 a multiplier, 4103 a half-time-slot (T/2) delay element, 4104 a low-pass filter, and 4105 an output terminal. Unlike the first conventional example, the carrier is rotated by 90 in phase by the half-time-slot delay element 4103. As with the first conventional example, consideration will ~e given as to how a detected output sisnal behaves in the presence of a two-wave multipath transmission line having a propagation delay time difference t which is not negligible as compared with the time slot.

FIG. 4~ is a diagram explanating how the detected output signal behaves when the transmitted signal as shown in FIG. 40 is subjected to the differential detection in the presence of the two-wave multipath propagation. FIG. 42(a) shows a phase transition of the D-wave. A phase transition of the U-wave which arrives with the propagation ~Z7~6~38 delay time difference T from the D-wave is illustrated in FIG: 42tb). The detected output at a particular time is the vector inner product of the combined phase of the two waves at that time and a phase attained by rotating the combined phase of the two waves in a ~eding half time slot by 90.Eor example, the detected output in a region B in FIG. 42(c) is the vector inner product of the combined phase of the two waves at a time B and a phase attained by rotating the phase at a time B' by 90.
FIG. 43 illustrates the combined phase of the D-and V-waves in order to determine the detected output at each of the times A through E. The ratio of the amplitudes of the D- and ~-waves is indicated by p, and the phase difference therebetween by a . For example, the absolute value of the detected output at the time B is the square of the line segment O~ in view of the fact that the vector OB is pe~ ciær to thevector OB' in FIG. 43. Therefore, the detected outputs at the respective times A through E in FIG: 42(c) are given, using the cosine theorem, as follows:
A ... indefinite ... an~l + p2 ~ 2p co~ a1 C ... an~1 + p2 + 2anP sin a) D ... an(1 + p2 + 2p cos a) E ... indefinite where an ~an ~ + 1) is a data sequence being transmitted.
In the regions A and E, the detected outputs are indefinite dep~ing on the data values in the preceding and ~27~88 subsequent time slots. A~tually, the cut-off frequency of the low-pass filter 4104 is selected to be low enough to prevent intersymbol interference. Therefore, the output signal that has passed through the low-pass filter 4104 is produced by filtering the'solid-line waveform of'FIG.
42tc), and constitutes a portion of an eye pattern as shown by the dotted line in FIG. 42(c). Since the regions B, D and the region ~ produce complementary detected outputs as described above, the eye will not be closed.
Furthermore, inasmuch as at least one of these effective detected outputs does not become smaller than the ineffective detected output in the region A or E, any fluctuation of the eye Ln the direction of the time base i9 reduced,'and any degradation of the bit error rate due to a follow-up failure of a re~x~w~ing clock is small.
With the second conventional digital signal transmlssion method, as described above, the bit error rate characteristics are largely improved in a mult~path fading tra~smLssion line by a kind of di~ersity effect by combining the mutually different outputs in the regions B, D and the re~ion C, making high-speed digital tran3mission possible (See, for example, "A Modulation Technique Suffering from Less ~iming Fluetuation of Eye Patterns in Multipath Fading" by S. Ariyavisitakul et al, Papers of Tech. ~x~p on ~un. Syst., l~ J~PA~, CS84-67, 1984).
Because the second conventional digital signal transmission method has two phase transitionc per time _ g _ ~27~i88 slot, however, it occupies a frequency bandwidth about ~wic~
that of the ordinary phase modulation, and hence is poor in frequency utilization efficiency. This method is, in principle, capable of only binary transmission, and cannot reduce the bandwidth through multivalued transmission- The bandwidth could be reduced to a certain extent by reducin the amount of phase transition or smoothing the phase transition more than a stepped pattern, but the bit error rate characteristics would be highly degraded ("A Further Study of Anti-Multipath Modulation Technique DSK ~
Analysis of Generalized DSK Modulation and Considerations for a Narrow-Band Scheme" by S. Ariyavisitakul et al, Papers of Tech. Group on Commun. Syst., IECE JAPAN, CS85-108, 1985).
Furthermore, according to the second conventional digital signal transmission method, ~he bit error rate characteristics can no longer be improved in principle when the delay time difference exceeds 0.5 in terms of t/T which is normalized with the time slot interval T. This is because in a region in which t/T is 0.5 or more, the regions B and D disappear, and the diversity effect which would be produced by combining the two kinds of detected outputs is no longer obtained.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide superior bit error rate characteristics in the presence of multipath fading and to improve frequency utilization efficiency through multi-phase transmission, or to improve bit error rate characteristics with respect to a multipath transmission line having a greater delay time difference for allowing high speed transmission of a digital signal.
~ In order to achieve the object, the present invention employs, in a transmission apparatus for transmitting digital data, a transmission signal having a lZ76688 phase transition waveform in each of the tlmc slot~s of data, the time-slot phase transition waveform having var~ing change rate of phase or a phase jump, a time-sl~t phase transition waveform in any desired time s]ot an~ a time-slot phase transition waveform in a time slot which comes in prescribed time slots subsequent to the de~ired time slot being identical to each other in shape irrespective of information to be transmitted, the information to be transmitted being present in a phase difference betwcen the same positions of the phase transition waveforms in the time slots that are spaced apart from each other by the prescribed time slots.
By using the above transmitted signal and effecting a differential detection employing a delay element capable of delaying the signal for the prescribed time slots, different detected outputs according to multipath propagation are produced. Through a kind of diversity effect produced by combining these outputs with a low-pass filter, bit error rate characteristics are highly improved in the presence of multipath fading for allowing high-speed transmission of a digital signal. By making the phase difference transmitting the information multiphase, a multivalued transmission is rendered possible with ease for increasing frequency utilization efficiency without degrading the characteristics with respect to the multipath propagation. Dependent on the kinds of phase transition waveforms in the time slots, improvement can be also attained for a delay wave with t/T > 0.5.
The above and other objects, features and advantages of the present invention will become more apparent from the following description taken in conjunction with the accompanying drawings in which preferred embodiments of the present invention are shown by way of illustrative example.

12766~38 DESCRIPTION OF THE PREFERRED EMBODIMENTS
A digital signal transmission method according to an embodiment of the present invention will hereinafter be described with reference to the drawings.
FIG. 1 is a phase transition waveform diagram showing an example of a phase transition waveform of a transmission signal transmitted by a digital signal transmission method according to the present invention:
A time-slot phase transition wavefonm ~(t) (0 < t < T) in a time slot of data is different from a conventional phase modulation method of a fixed value in that its differential coefficient is variable, or it has a phase jump. FIG. 1 shows an example in which there is a phase jump. The time-slot phase transition waveform ~(t) has a stepped pattern with a phase jump, indicated by , between a front halE portion Tl and a rear half portion T2. 'rhe phase transition waveforms in a first time slot and a ~n ~ l)th time slot which are spaced apart from each other by n time slots are identical to each other in shape, and.are shifted by ~ according to the.l-n~ormation to be transmitted. Stated otherwise, n-time-910t differential coding is effected. For example, when a binary-phase qystem with 0 and ~ as 9 is employed, information of one bit per time slot.can be transmitted, and when a quadrature-phase system with 0, ~/2, ~, 3 ~/2 as 3 is employed, information of two bits per time slot can be transmitted.
is generally indicated as follows:
~ 2m (m = 2P, p - 1, 2, 3...) ...~1) where i is the data value which is converted into a Gray code to be transmitted, and 0 < i < m, i ~ Integer.
Therefore, if the phase transition waveform in the first time slot is ~(t~, then the phase transition waveform in the ~n + l)th time slot i~ expressed as ~(t - nT) I ~.

~Z766~38 Assuming that a phase shift bearing information is indicated as a phase shift ~a(t) from the absolute phase, the phase shift ~a(t) is a stepped function having a fixed value in each time slot, and can be expressed, using a data vzlue sequence idq produced by the n-time-slot differential coding of a data sequence iq ~q ~ Integer) that has been converted to a Gray code to be transmitted, as follows:
a~
~t) = q~ ~ idq 2m {U~t-qT) - U(t - ~q-l)T)}

I l (t > O) U(t) -0 (t < O) ...~2) There may be a plurality of kinds of t~Y~slot phase transition waveforms ~(t). For the n-time-slot diferential coding, n kin~s of time-slot phase transition waveforms ~l(t), ..., ~n(t) are available.
If it is assumed that ~ r(t) - O (t < O, t > T, 1 ~ r ~ n) ...(3) then the general formula of~the phase transition waveform ~(t) of a transmission signal in the digital signal transmission method of the present invention can ~e expressed, using the equation ~2), as follows:
~t) ~ r (t - (qn+r-l)T) + ~ (t) q=-~ r=l _ ~ n~ ~r (t - (qn+r-l)T) r=l ~ 2 + q=_x idq ' -m {U(t-qT) - U~t-(q-l)T)~ (4) The phase transition waveform of the transmission signal according to the present invention resides in the first term of the equation (4), with the second term being the same as that of the conventional differential coding phase modulation.
The time-splot phase transition waveforms ~l(t), ~2(t), ~n(t) may include those which are ` 127~688 identical to each other. In a special case, all of them may be identical to each other. At any rate, the ti~e-slot phase transition waveforms ~t~ that are spaced from each other by the n time slots need to be identical to each other.
~he value of n may be 1, and in such a case, the t~me-slot phase transition waveforms ~(t) are of one kind, and the phase transitiOn wave Frms in all of the time slots are of identical shape. Where the time slot pnase transition waveforms ~t) are of one kind, the phase transition waveform ~(t) of the transmission signal is expressea, using the equation (2), as follows:

~(t) = ~ ~(t-qT) + a (t) q co .
= ~ ~tt-qT) q=_~

~ q~ ~ idq- 2m {U(t-qT) - U(t-(q-l)T)}

FIG. 1 shows the stepped tlme-slot phase transition waveform ~t) as described above.
A duty ratio ~l/T indicating the position of a phase jump may be of any desired value. The duty ratio may of course be SO %, that is, the phase jump ~ may be located centrally in the tirne slot. The phase jump ~ may be of any desired magnitude and may be leading or lagging in phase. In a special case, the phase jump ~ may be ~. As described later on, as the phase jump ~ is larger, i.e., it is closer to ~, the characteristics under multipath fading are improved, but the envelope with a limited band varies to a - `
large extent.
As described above, there may be a plurality of kinds of time-slot phase transition waveforms ~(t)o~IG~ 2 shows a case in which there are a plurality of kinds of phase jumpC ~. FIG. 3 shows a case in which there are a plurality of kinds of duty ratios. FIG. 4 illustrates a case in which there are a plurality of k~s of ccmb~tions of phase jumps .. . ~ . ~ . .

~27668E~

and duty ratios. FIG. 5 shcws a case in which the transition dire~tions of p~ase jumps ~ are alternately leading and lagging. FIG. G illustrate~
a case Ln which there are a plurality of kinds of masnitudes of phase jumps ~, and FIGS. 7 an~ 8 show cases in which there are . a plurality of kinds of duty ratios, with the distance n between oorresponding time slots being of an even numker and in~icated by 2n'. FIG. 9 shows a case in which the transition directions of phase jumps ~ constitute a quasi-random binary seguence having a period equal to the di~tance n between oorresponding time slots~ FIG. 10 shows a case in which there are fur~her a pl~x~ity of k~s of magnitu~es of phase jumps ~, and FIGS. 11 and 12 show cases Ln which there are a plurality of kinds o duty ratios.
FXG. 13 is a phase transition waveform diagram showing the phase transition waveform of a transmission signal transmitted by a digital signal transmission method of the present invention, in which the time-slot phase transition waveforms ~(t) are of one kind and of a stepped pattern with the duty ratio being 50 % and the phase jump ~ - ~, and in which n = 1, i.e., one time-510t dlfferential coding, and multiphase number m - 4, so that two bits can be transmitted per time slot.
The time-slot phase transition waveforms ~(t) may have a plurality of steps as shown in FIG. 14.
The time-slot phase transition waveforms ~(t) may be of a triangular shape as shown in FIG. 15, instead of the stepped pattern. The gradient of straight lines constituting the phase transition waveform corresponds to a frequen~y shift from the carrier frequency. Therefore, for ~276688 the e~ample of FIG. 15, one time slot is composed of two fre~uencies. The gradient of straight 1ines cons~ituting the time-slot phase transition waveform may be of various ~inds as shown in FIG. 16, and may have a p`nase jump as sho~n in FIG. 17. Fur~herm~ore, in orcer to reduce the transmission spectral wid~h, the phase transitlon waveform may be of a pattern p.ocuced by smoothins any one of the above-mentioned ~aveforms. For e.ca~ple, it may be of a raised cosine waveform or a Gaussian waverorm as shown in FIG. 18.
A method of producing the transmission signal as described above will hereinafter be descrioed with reference to embodiments of t~e invention.
FIG. 19 is a block diagram of a generator circuit for a transmission signal according to a first embodiment of the present invention. Designated in FIG. 19 at 1901 is a data input terminal; element 1902 is d difrecential coding circuit, element 1903 is an oscillator; element 1904 is a waveform yenerator circuit;
ëlernent 1905 is a quadratucelnodulator, and element 1906 is a transmission signal output terminal. Digital data to be trarl~mitted is supplied from the data input terminal 1901 and subjected to differential coding in the differential coding circuit 1902. The waveform generator circuit 1904 generates modulation signals in I- and Q-phases according to the differential coded data- The oscillator 1903 produces a ~ carrier, which is modulated by the modulation signals in the I- and Q-phases in the quadrature modulator 1905 into a i8~

transmission signal that is outputted from the transmission signal output terminal 1906.

FIG. 20 shows a circuit arrangement of the quadrature modulator 1905 shown in FIG. 19. Denoted in FIG.
20 at 2001 is a 90 phase shifter; elements 2002 and 2003 are balanced modulators, and element 2004 is a combiner.
The carrier signal supplied from the oscillator 1903 is modulated by the I-phase modulation signal from the waveform generator circuit 1904, using the balanced modulator 2002, into an I-phase modulation signal from the waveform generator circuit 1904, using the balanced modulator 2002, into an I-phase modulated signal. The carrier signal is also phase-shifted by 90 by the 90 phase shifter 2001, and then modulated by the Q-phase modulation signal from the waveform generator circuit 1904, usiny the balanced modulator 2003, into a Q-phase modulated signal. The I- and Q-phase modulated signals thus produced are combined by the combiner 2004 into a modulated transmission signal, which is outputted from the transmission signal output terminal 1906.

FIG. 21 illustrates a circuit arrangement cf the diff~rential coding circuit 1902 shown in FIG. 19. Denoted at 2101 and 2104 are Gray code converter circuits; element 2102 is an adder, and element 2103 a delay element.
When the multiphase number is m (m = 2, 4, 8 ... ), a p-bit parallel data value sequence indicated by the equation (1) is applied to the Gray code converter circuit 2101. The data value sequence iq that has been converted ~ into a Gray code is applied to the adder lZ76688 2102 in which it is added (MOD m) to data which has been produced by delaying the output of ~he alder ~02 by the delay element 2103 for the n time slots, i.e., n clock ulses. The output from the adder 2102 is converted by the Gray code converter circuit 2104. Therefore, the p-bit input parallel data value sequence is converted into a Gray code to produce a p-bit parallel data value sequence idg subjected to n time-slot differential coding.

FIG. 22 is illustrative of a circuit arrangement of the waveform generator circuit 1904 with respect to a quad-rature-phase system in which the phase transition waveform ~(t) is indicated by the equation (5). Designated at 2201 is an I-phase data input terminal; element 2202 is a data clock output terminal; element 2203 is a Q-phase data input terminal; elements 2204 and 2206 are shift registers;
element 2205 is a binary counter; element 2207 is a ROM
(Read-Only Memory); element 2208 is a clock generator;
elements 2209 and 2210 are D/A converters; elements 2211 and 2212 are low-pass filters; element 2213 is an I-phase modulation output terminal, and element 2204 is a Q-phase modulation output terminal. For the ~uadrature-phase system, the output idg from the differential coding circuit 1902 is a 2-bit parallel data with its most- and lesat-significant bits being supplied from the I- and Q-phase data input terminals 2201 and 2203, respectively.
The supplied data sequences are delayed respectively by the shift registers 2204 and 2206 to provide modulating data in the present time slot and modulating data in time slots subsequent to and preceding in the present time slot. In FIG. 22, the shift registers 2204 and 2206 produce 12766~38 at Qd the modulating data in the present time slot, and at Qe through Qg and Qa ~h~gh Qc the mo~ating da~a in the ~hxe time slots ~ubsequent to and preceding the present time slot.
The I- and Q-phase modulating waveforms are written in the ROM 2207 according to the mcdulating data. In FIG. 22, each time slot i8 composed of 16 ~ampling point~. Addresse~ A7 through A17 of the ROM 2207 are used as select signals for determining which modulating data is to be det~cted, and are supplied with the mXh~ating data in the three time ~lot~
sub~equent to and preceding the present time ~lot. To addres~es A0 through A3 of the ROM 2207, there is ~pplied a signal, as a modulating-waveform readout signal, which is produced by Ere~uency-dividing a reference clock slgnal from the clo~k generator 2208 with the binary counter 2205.
The ROM 2207 produces outputs X0 through X7 and Y0 through Y7 which are converted into analo~ signals by the D/A
c~nverter~ 2209 And 2210' and the low-pa8s filters ~211, and 2212 which remove olded components, the analog signals serving as the I- and Q-phase modulating signals. For modulation in a re nultl-pbase system such a3 an octal-phase system, it 18 neCe~sary ~o-ha~e a~ many shift registers as the number o~ p in the equatlon (1) and corresponding ROM addresses.
The modulatLng waveform in each time slot which is written in the ROM 2207 will be described below.
Basically, the I- and Q-phase modulating waveforms MI(t), MQtt) may be obtained by the following equations from the pha~e transition waveform ~t) of the transmission signal ~27668~3 which is derived by the equation (5) from the data value sequence idq subjected to the differential codingand to be transmitted:
. MI(t) = cos ~tt) -' MQ(t) = sin ~(t) ...(6) Since these waveforms are wideo2nd signals as they are, they are limit~?d in band width by a oand~ itin~ filt~ havinq an impulse response h(t). The equations (6) are then modified as follows:

+to MI(t) = Jcos ~(t ~ h(~)dl ...(7) r~to MQ(t) =) sin ~t -1) h~T)d~ ' Various band-limiting filters capable of passing lower frequencies, such as of the cosine-square type or the Gaussian type, may be employed. Such various filters have different impulse responses h~t). As an example, the impulse response h~t) of a cosine-square type filter having, a cut-off angular frequency ~o and a roll-off coefficient y is given by: ' ~

' h(t) = ~ sin ~ot cosY ~o t ...(8) ~ ~ot 1-(2y ~o t/~ )2 The cut-off angular frequency ~o should be selected to be about the same as the modulation rate which is a reciprocal of the time slot interval T. - :
In the R0.~ 2207 shown in FIG. 22, there are written the I- and Q-phase modulating waveforms MI(t) and MQ(t) , :, . : '.' .
- 20 - .

1;2766~1~

for one time slot according to the equations (7). The range of integration (-to, to) in the equations (7) is selected to be about the same as the spread of the im?ulse response h(t), and is equal to the three time slots subsequent to and preceding the pzesent time slot in the arranqement OL FIG. 22. The mcdula-~Lng dat~-in the chree time slots subsequent to and preceaing the present .ime slot are required in order to calculate the phase transition waveform ~(t) from the equation (5). Therefore, the waveform dat~ fo~ one tim~ slot is calculated by the ~uations (7) with respect to all of the ~lo~ul~ting datd patte~ls in the ~e-ent ti.ne slot & the three time slots ~ sequent to and preceding the present ti~e slot, and is written in the RCL~ 2207. A modulating waveform is selected by the addresses ~4 through A17 of the ROM 2207 which are the modulating data in the present time slot and the three time slots subsequent to and preceding the present time slot.
This is substantially the case with a plurality of kinds of phase transition waveforms ~(t) in the time slots just like the time-slot phase transition waveform ~(t) is indicated by the equation (4). The I- and Q-phase modulating waveforms MI~t), MI(t) for one time slot, as calculated by the equations (7), may be written in the ROM.
In determining ~(t) in the equations (7) from the equation 14), it is necessary to find the value of r (1 < r < n) in the time-slot phase transition wavefon~ ~r(t) in the present time slot. Therefore, the waveform datas to be written in the ~276~
ROM are calculated with respect to not only the modulating data patterns but also representative of which one is the time-slot phase transition waveform ~r(t) in the present time slot.
Accordingly, the circuit arranfgement of the waveform generator 1904 shown in FIG. 19 must be modified as shown in FIG. 23. Denoted in FIG. 23 at 2201 is an I-phase data input terminal; element 2202 is a data clock output terminal; element 2203 is a Q-phase data input terminal;
elements 2204 and 2206 are shift registers; element 2205 is a binary counter; element 2208 is a clock generator; element 2209 and 2210 are D/A converters; elements 2~11 and 2212 are low-pass filtersi element 2213 is an I-phase modulation output terminal, and element 2214 is a Q-phase modulation output terminal. These components are the same as those shown in FIG. 22. The circuit arrangement of FIG. 23 differs from that of FIG. 22 in that a binary counter 2301 indicating the present value or r is added, and addresses A18 and A19 are added to a ROM 2302 for selecting a waveform according the value of r. The binary counter 2301 has a period n, which is 4 in the arrangement of FIG. 23.
A method of detecting a transmission signal in the digital signal transmission method of the present invention, as described above, will be described belo~.
In the digital signal transmission method of present invention, a signal is detected by a differential detector having a delay line for n time slots. This detect-ing method is the same as the conventional method, and described in detail, for example, in "Data Transmission", by William R. Bennet ~ James R. Davey, McGraw-Hill Book Co. New York, 1965. The detecting method will be described briefly hereinbelow.

-127~i688 FIG.24 illustrates a circuit arrangement of adifferential detector of a binary-phase system. Designated in FIG. 24 at 2401 is an input terminal, 2402 a multiplier, 2403 a low-pass filter, 2404 an n-time-slot (nT) delay element, and 2405 an output terminal. The signal is delayed for the n time slots by the n-time-slot delay element 2404.
Unlike the second conventional arrangement, the carrier phase remains identical at the input and output. The low-pass filter 2403 not only removes a component having a frequency which is twice as high as that of the carrier produced by the multiplier 2402, but also serves to combine a plurality of detected outputs. The low-pass filter 2403 is preferably a so-called Nyquist filter naving a cut-off frequency that is half the symbol transmission rate, i.e., (1/2)T, and also having attenuation characteristics asymmetrical with respect to the cut-off frequency.

FIG. 25 shows a circuit arrangement of a dif-ferential detector of a quadrature-phase system. Denoted in FIG. 25 at 2501 in an input terminal; elements 2502 and 2506 are multipliers; element 2503 is a -45 phase shifter;
element 2505 is a +45 phase shifter; element 2504 is an n-time-slot (nTJ delay element; elements 2507 and 2508 are low-pass ~ilters; element 2509 is an output terminal A, and element 2510 is an element ~2~6~i88 output terminal B. The circuit arrangement of FIG. 25 differs fro~ that of FIG. 24 in that time -45 phase shifter 2503 and the +45 phase shifter 2505 are employed to effect differential detection with respect te two mutually perpendicular axes for demodulating 2-bit parallel data.
The other operation is the same as that of the circuit arrangement of FIG. 24.

FIG. 26 shows a circuit arrangement of a differential detector of an octal-phase system. Denoted in FIG. 26 at 2601 is an input terminal; elements 2602 through 2605 are multipliers; element 2606 is an n-time-slot (nT) phase shifter; element 2607 is a -22.5 phase shifter;
element 2608 is a +22.5 phase shifter; element 2609 is a +67.5 phase shifter; element 2610 is a -67.5 phase shifter; element 2611 through 2614 are low-pass filters;
element 2615 is a comparator; element 2616 is an output terminal B. The phase shi~ters 2607 through 2610 serve to effect differential detection with respect to three axes which are 45 displaced for demodulating 3-bit parallel data. The comparator 2615 detects whether or not both inputs applied thereto are of teh same polarity.

~ The manner in which the digital signal transmission method of the present invention exhibits good bit error rate characteristics in the presence of multipath fading will be described below.

FIG. 27 is a diagram in which a transmission system of a binary-phase system having stepped time-slot phase transition waveforms as shown in FIG. 1 is ~Z766~3~

employed as a transmission signal to be transmitted by the digital signal transmission method of the invention, and explains a detected output signal produced by detecting such a transmission signal with the differential det~ctor of FIG. 24. FIG. 27(a) shows the manner of D-wave phase transition of a time slot and a time slot which is n time slots subsequent to the former time slot. The t~æ-slot phase transition waveforms in both o th~ time slots are of an identical stepped pattern as described above.
The phase transision of the U-wave which comes with a time delay equal to the propagation delay time difference ~ that ~ s not negligible as compared with the time slot interval T
is as shown ~n FIG. 27~b). The detected output at a particular time is the vector inner product of the combined phase of two waves at that time and the combined phase of the two waves in a time ~hich is the n time slots prior to the above time~ For example, the detected output in a region B in FIG. 27~c) is the vector inner product of the combined phase of the two waves at a time B' and that at a time B.
FIG. 28 shows the combined phase of the D- and U-waves in order to determine the detected output at each of the times A through E. The ratio of the amplitudes of the ~- and U-waves is indicated by p, and the phase of the carrier of the U-wave as seen from the carrier of the D-wave is indicated by a. The phase axis of FIGS. 27~a) and 27(b) is such that the leading direction is positive.

lZ76~88 Based on FIG. 28, when the waveform is not dis.o ed bv the low-2ass filter 2403, or the cut-off freauenct is sufficiently high as compar-d with the ~ata t 2nsmission rate, the detected outputs at the res~ec_iv2 times A
through E in FIG. 27(c) are as follows:
~egions A, E
inde inite Regions B, D
1 + p2 + 2P cos ~ ... (9) Region C
1 + p2 + 2 Pcos(~
In the regions A and E, the detected outputs are rendered ind~finite depending on the data values in the preceding and subsequent time slots. Actually, the cut-off frequency of the low-pass filter 24~3 is selected to be low enough to prevent intersymbol interference. Therefore, the output signal that has passed through the low-pass filter 2403 is proauced ~y filtering the solid-line waveform of FIG.
27(c), and constitutes a portion of an eye pattern as shown by the dotted line in FIG. 27(c). Since the regions B, D and the regions C produce complementary detected outputs and will not be simultaneously eliminated with respect to any value ofp or ~, the eye will not be closed.
Furthermore, inasmuch as at least one of these effective detected outputs does not become smaller than the ineffective detected output in the region A or E, any fluctuation of the eye in the direction of the time base is ~766~8 reduced, and any degradation of the bit error rate,due to a follow-up failure of a re~x~ucing clock i~ small.
Detected outputs under two-wave multipath propa~ation, includln~ a multiphase system, will be considered. Assuming that a transmission data sequence is an tan = + 1), a multiphase number i~ m (m - 2, 4, 8, ...), and complex m~ltiplicative noises indicating fading of the D- and U-waves are sl(t~' ant ~2(t) the detected outputs in the regions B, C, D ca~ be expressed as follows tthe complex envelopes of the D- and U-waves are expressed by sl~v, s2-v where v (¦v¦ - i) is the complex envelope of a transmitted wave):
Regions B, D
an5in~ ~/m) ~lsl ~ s212) Region C ... ~10) an~in~ ~/m~ ~Isl exp~ 52l2~
It should be noted in the equations ~10) that the carrier phase of the D-wave in the detected output of the region C
is shifted by ~.
A~ described above, the principle of the present invention for improving a digital signal transmis-~ion in the presence of multipath fading can be said to be a kind of diversity effect by combining different detected outputs. Therefore, an appropriate diversity model can be considered for evaluating bit error rate characteristics. Now, a bit error rate in the presence of two-wave multipath fading which meets the condition of a delay time difference T with the both regions B, D and ~2~7~688 region C being present will be evaluated.
A model of a maximum ratio combiner 2901 supplied with, as two branch inputs, signals Ul~v~ u2-v tv is the complex envelope of a modulated signal, ¦v¦ = 1) accompanied by fading expressed by complex mNltiplicative noises ul(t), u2(t), and a differential detector 2902, as shown in FIG. 29, wdll be concldered. In the maximum ratio combiner 2901, the sum of S/N
ratios of the inputs is equal to the S/N ratio of the output ~where N is noise power). Therefore, at this time, the complex envelope of the received signal at the input of tne differential detector 2902 is indicated by . _ ~¦ul¦2 + ¦u2¦2 v, and the detected output from the differential detector 2902 is given by:
an(¦ull2 1 lu212) ...~11) When the two detected outputs in thè equation (10) are combined equally, the first and second terms of the equation ~11) correspond to the two detected outputs of the e~uation (10), and are expressed by:
Ul ' ~sin( ~m)-(sl + s2) ...~12) ul ~ ~sin( ~/m)-tsl exp~ s2) Thus, lt will be understood that FIG. 29 shows a diversity model employing stepped time-slot phase transition wave-forms in the digital signal transmission method of the present invention in the presence of two-wave multipath fading.
The equations (12) indicate that the diversity effect, strictly speaking, has, at diversity branches, a combined wave of coming waves having the delay time di~ference T and ~2~;688 a combined wave of those coming waves with the carrie~
phase difference ~etween them being shifted by ~.
The diversity characteristics employing the maximum ratio combiner can analytically be evaluat~d.
Assuming that R is a complex variance matri~ < ~ > (i, j = 1, 2) of {ui/~N } (i = 1, 2), the average bit error rate Pe when the S/N ratio is large is expressed by the following equation, as desc~ibed in "Communication Sys,ems and Techniques", by M. Schwartz and W. R. Bennett & S.
Stein, McGraw-Hill Book Company, New York, 1966:

2-ae~ R ...(13) Assuming that the fading of the D-wave and the fading of the U-wave are independent of each other an~ the average D~U ratio is OdB, det R is calculated as follows:
det R = ~`f sin( ~/m)-sin~ /2)}2 ... (l~) Y = S/N ratio Therefore, the average bit error rate Pe is:

_ l _ ... (15) 2 {y sin( ~/m) sin(~ /2)}2 Thus, from the viewpoint of the bit error rate, the optimum value of ~ is ~, and the average bit error rate is degraded twice at ~/2 and seven times at ~/4. With respect to the effect on the multiphase number m, the average bit error rate for q~ture-phase syste~ is degraded ~ice that for binary-phase sys~
and the average bit error rate for octal-phase syste~ is degraded seven times that for bin2ry-phase system. According to the paper "A Further Study of Anti-l~ultipath Modulation Technique DSX

. . .; .

¢~

i~766~38 ~ Analysis of Generalized DSX Modulation and Considera-tions for a Narrow sand Scneme" by S. Ariyavisitakul et al, Papers of ~ h. Grou? cn C~n. Syst., IEC~ ~P~, CS35-103, 193,, th.e a~-er2ce bit ~rror rate Pe of 9-DS~ w`nich is calculated bv th~ same method and is a generalization of DSX discussed as th~ second conventicn l dicit~l sicnal ~rans~ission m~thod is e.~ressed by:

1 ...(16) Pe 3S:~ 2- {y sin a - sin( ~/2)}2 Comparison of the equations (15), (16) shows that the former is different from the latte~ in that the parameter which dc~inates the diversity effect is separated from the amoun~ of phase transition ~/m(a) for delivering information. Therefore, by selecting ~ to be of an optimum value ~, thR digital signal transmission method of the present invention can produce better characteristics than those of the conventional digital signal transmission methods, particularly DS~, i.e., the second conventional arrangement. For e~ample, the quadrature-phase digital signal tra~smission method with ~ = ~ of this in~ention has bit error rate characteris.ics equal to those of ~/2-DS~ which is the secor.d conventional digi~ signal transmission method, and has frequency utilization efficiency about twice that of the conventional method since the method of this invention is capable of quadrature-phase transmission while the conventional method is capable of only binar~y-phase t~ansmiS-sion. Thus, the digital signal transmission method of this invention has better bit error rate characteristics .... . . ....

127~

and it is capable of multiphase transmission.
The aforesaid analysis has been made with refer2nce to two-r~ave multi?a_~ faaing, bu. ca~ easily be-eYtendec to multiwaves. For e.Yam~le, when three waves ar-involved, there are three detec.ed out?uts, and the model shown in EIG. 29 has three branches, resulting in much be~-er characteristics. Diversity branches are also inc-eased whe~ a pluralit~ o. waverorms are employed as t~-slot phase t-ansition waveforms ~(t).
While the ~res~nt ir.v~ntion has ~e~n des~ribe~ above~th ste?ped wavefor~s e.~?loyed as the ti~e-slot phase transition wavefor~.s ~t), it is easily e:~tended to any desired dif~erPn. waveforms.
FIG. 30 is ~ diagram illustfative of a detected output in the presenc~ of two-wavelnultipath ~ropayation, as with FIG. 27, with respect to any time-slot phase transition waveforms ~(t)- Like the exm2?1e of FIG. 27, the detected output is roughly classified into three regions F, G, H, the regions A, E corresponding to the regions F, H, respectively. In these regions, the output is dn ineffective detected output and has a polarity which does not necessarily coincide with the value of data transmitted. The regions B, C, D correspond to the region G. In this region, t~e output is an effective detected output and has a polarity WhiCh necessarily coincides with the , . -- 31 --` . . ` . . ` :.. ~?: , ~2 ;t668~ -value of data transmitted. Although there is no de~ini,e divisions in the region G, different types of detecled outputs appear as indicated by the solid line in FIG.

30(c). The wavefo-m shown by the solid line in FIG. 30(c) is filtered by a low-pass filter to form a portion o- an eye pa'_ern as indic~ted by the dot~ed line in FIG. 30 ( c) .
The detected out?ut in the resion G can be expressed, with z used as a parameter, similarly to the ecuation (10), as follows:
Resion G
. an sin( ~/m) ( 15l ex?{; II)(Z)} + 5~ e~:?{i 'r~(Z ~ 'r)~ 12) = an sin( ~/m) ;(¦5l e~p[j { ~(z) - ~(z -.)~] + s2l2) where T < z < T ...~17) Thus, inso~ar as the time-slot phase transition ~Javeforms ~(t), which meet the condition:
~ (z) - ~(z - T), cons,. (T < z < T) ...(18) the equation (17) is not constant, and it can be seen that bit error rate characteristics in the presence of multipath fading can be improved by the kind of diversity effect produced by combining different detected outputs. The conditio~ of the equation (18) indicates that the time-slot phase transition waveforms ~(t) have a varying rate of change of phase or a phase jump.
When the raised cosine waveform or Gaussian waveform as shown in ~IG. 1~ is employed, the spectrum utill2ation ~ 27~68~3 efficiency is good stated otherwise, greater protection against band limitations i9 obtained.
Examples of the average bit error rate characteristics in the presence of two-wave Rayleigh fading having a delay time difference will be described with reference to typical examples of the digital signal transmission method of this invention.
FIG. 31 is a graph in which the.time-slot phase transition waveforms are of one type in the form of a stepped pattern having one phase jump ~ ) at the center of the time slots as shown in FIG. 1, the graph ~howing average bit error rate characteristics for the quad-rature-phase system with respect to the S/N ratio. The average bit error rate characteris~ics o the con~entional digital signal transmission method, QPSK (Quadrature Phase Shift Reying), are al~o shown in the same graph for comparison. As shown in FIG. 31, in QPSK, errors are produced that cannot be reduced even by increasing the S/N ratio. According to the digital signal transmis6ion method of this invention, no such phenomenon occur~, and the bit error rate characteristic~ are highly improved.
FIG. 32 similarly shows an average bit error rate with re~pect to the delay time difference T. In the range ofO~T/T < 0.5, the average bit error rate is improved to a great extent. At T~T = 0 or T~T 2 0.5, no improvement is achleved, and the characteristics are closer to those of - 33 ,~

lZ~66~38 .

QPS~ because the region C or the resions s, D dis~?~e~r in FIG. 17, losing the diversity effect.
The bit error rate can be improved wi~h respec~
to a la-se value o. ~/T by s~iftinc the posi~~on o ~.~e phase 3ump ~ from the center of the time slcts. ~IG. 33 shows an average bit error rzte with res?ec- tO ~he aelav tim_ di ~erence T when the phase jum? ~ is pos .ioned at an in.ernallv dividea point in the ra,io 1:2. Tne bit error rate can be im?roved thereby up to the range of about T/T <
0.7. ~ DSX discussed as the second conventional method, it is im?ossible in principle to im?rove the bit error rate at ~/T > 0.5.
FIGS. 34 and 35 show averaqe bit error rate characteristics for the quadr2ture-phase system with respect to the delay time difference T with the time-slot phase transition waveforms being triangular and raised cosine waveforms as shown in FIGS. 15 and 18. The bit error rates are similarly improved up to about ~/T = 0.7.
In each of FIGS. 34 and 35, the maximum value of phase transition in the time slots is ~, and the characteristics are degraded below this value and are not virtually improved above this value.
In this embodiment, as describnd bove, the bit error rate characteristics can be improved to a large extent in the presence of multipath fading by employing a transmission signal having a phase transition waveform ineach of time slots of data, the time-slot phase transitio:- wav~form having .. ..

: ..

~ ~:76688 a va~ying rate of change of phase or a phase jump, the time-slot phase transition waveform in any desired time slot and the time-slot phase transition waveform in a time slot which comes prescribed time slots subsequent to the desired time slot being identical to each otherin shape irrespective of information to be transmitted, the information to be transmitted being present in a phase difference between same positions of the phase transition waveforms in the time slots that are spaced apart from each other by the pr-escribed time slots. The spectrum utilization efficiency is increased by multiphaes transmission as compared with DSK
described above as the second conventional method. By selecting time-slot phase transition waveforms, bit error rate characteristics can be improved even with respect to multipath fading having a larger delay time difference.

A second em~odiment of the present invention will hereinafter be described with respect to FIG. 36.

FIG. 36 shows a transmitter circuit in a digital signal transmission method according to a second embodiment of the present invention. Denoted in FIG.36 at 1901 is a data~input terminal; element 3603 is a transmission signal generator circuit; the data input terminal 1901 and the transmission signal generator circuit 3603 are identical to those in the first embodiment shown in FIG. 19; elements 3610 through 3612 are first through kth antennas of a k system; elements 3607 through 3609 are level adjusters of the k system, and elements 3604 througn 3606 are first through (k-1) delay elements of a k-1 ~ ~6688 system. The level adjusters 3607 through 3609 may also have an amplification capability. As a detecting method at the receiving end, n t~me-slot differential detec-tion as shown in FIGS. 24 through 26 according to the first embodiment is employed.
The digital signal transmission method according to the aforesaid arrangement will hereinafter be described with reference to FIGS. 27, 32 through 36, and the equation (18).
FIGS. 32 through 35 show bit error rate characteristics, as described above~ for the transmission signal which is produced by the transmission signal generator circuit 3603, propagated through two-wave Raileigh fading paths having the delay time difference 1, and then received and detected. Now, it is assumed that the delay time difference T of the propagation paths, i.e., delay spread, is smaller than the time slot interval T. This condition is applied to the case where the delay spread is small as in a building, or the transmission rate is low.

When T/T is close to 0, e.g., where the time-slot - phase transition waveforms ~(t) are of a stepped pattern, the region C in FIG. 27 is reduced. Where ~(t) is of a general waveform, the lefthand side of the equation t18) varies to a smaller extent as z varies. This results in elimination of the diversity effect which would be produced by combining different detected outputs as described with reference to the first embodiment. As a consequence, as .:,, . : ,.. .
_ ~ ~6 ~ _ ~ 276~i~38 T/T approaches 0 in FIGS. 32 through 35, the bit error rate characteristics are less improved. But, by giving, at the transmitting end, a delay in the range o~ from 0 to 0.5 or from 0 to 0.7, which is the improving range for ./T, the bit error rate characteristics can be im?roved.
The delay elements 3604 throuch 3606 shown in FIG. 36 serve to give such a delay in the transmitti~g end.
The delay elements should be designed to set, to Tm/T, the time difference rm between a wave arriving first and a wave arriving last at the receiving end, including a delay due to the difference in paths from the respective antennas, so that the delay will not exceed 0.5 or 0.7, which is the maximum improving range for r~T that is determined by the time-slot phase transition waveforms ~(t). The level adjusters 3607 through ~609 serve to substantially equalize, at the receiving point, the average levels of waves accompanied by fading and coming from the respective antennas. The first through kth antennas are required to be located at spaced intervals or be composed - `
of antennas having different planes of polarization such ... ~
- that the fadings of the respective paths from the antennas - .;
to the receivi`ng point are mutually uncorrelated. One simplest and most effective arrangement is the case in -which k = 2. In this case, the time difference Tm between ~
waves arriving from two antennas is preferably selected to -be 0.2 to 0.4 in terms of Tm/T, which is the best point of an bit error rate determined by the time-slot phase transition .... .
... . - . - . ~ . - ~ ~ _ 3 7 _ . . ; . ~

1~7~688 -waveform ~(t).
With the second embodiment of the present invention, as described above, one transmission signal is transmitted with a time delay from different antannas for thereby obtaining a diversity effect even when T/T is small, so that the bit error rate characteristics can be improved. The diversi~y is highly effective in reducing the size of a receiver and making it portable since only one antenna is required on the reciever.

.. ... ..

:, . .. ;
. .

~i~..r . .

Claims (22)

THE EMBODIMENTS OF THE INVENTION IN WHICH AN EXCLUSIVE
PROPERTY OR PRIVILEGE IS CLAIMED ARE DEFINED AS FOLLOWS:

1. A method of transmitting a digital signal, employing, in a transmission apparatus for transmitting digital data, a transmission signal having phase transition waveforms in respective continuous time slots of data, each phase transition waveform in each of the time slots having a varying rate of change of phase or at least one phase jump, wherein phase transition waveforms in any two time slots which are spaced apart from each other on a time axis by a prescribed number of time slots are identical to each other in shape irrespective of an information to be transmitted, the information to be transmitted being present in a phase difference between the phase transition waveforms in any two time slots spaced apart from each other on a time axis by the prescribed number of time slots.

2. A method according to claim 1, wherein the phase transition waveform in each of the time slots is a raised cosine waveform.

3. A method according to claim 1, wherein the phase transition waveform in each of the time slots is a Gaussian waveform.

4. A method according to claim 1, wherein said phase difference is any one of angles obtained by equally dividing it by a number which is a power of 2.

5. A method according to claim 1, wherein said transmission signal is delayed and transmitted from a plurality of different antennas each with a time delay.

6. A method according to claim 1, wherein said transmission signal is detected by differential detection using a delay line capable of producing a delay corresponding to said prescribed number of time slots.

7. A method according to claim 1, wherein said phase transition waveform in each of the time slots is of a stepped pattern having at least one phase jump.

8. A method according to claim 7, wherein said stepped pattern has a plurality of phase jumps.

9. A method according to claim 7, wherein said stepped pattern has a single phase jump.

10. A method according to claim 9, wherein the phase jump occurs at the same position in each of the time slots.

11. A method according to claim 10, wherein the phase jump occurs at a center position in each of the time slots.

12. A method according to claim 9, wherein the phase jump in each of the time slots occurs at one of a plurality of kinds of positions.

13. A method according to claim 9, wherein the amount of the phase jump is t.

14. A method according to claim 9, wherein the direction of the phase jump is the same which is leading or lagging, and the amount of the phase jump is the same, in all of the time slots.

15. A method according to claim 9, wherein the direction of the phase jump is the same which is leading or lagging in all of the time slots, and the amount of the phase in each of the time slots is one of a plurality of kinds of amounts.

16. A method according to claim 9, wherein the direction of the phase jump is alternately leading and lagg-ing in each adjacent two of the time slots, and the amount of the phase jump is the same in all of the time slots.

17. A method according to claim 9, wherein the direction of the phase jump is alternately leading and lagging in each adjacent two of the time slots, and the amount of the phase jump in each of the time slots is one of a plurality of kinds of amounts.

18. A method according to claim 1, wherein the phase transition waveform in each of the time slots is composed of a plurality of straight lines of different gradients.

19. A method according to claim 18, wherein said gradients are of two kinds.

20. A method according to claim 18, wherein said gradients, i.e., frequencies, are of three kinds or more.

21. A method according to claim 18, wherein said phase transition waveform in each of the time slots has a phase jump.

22. A method according to claim 18, wherein said phase transition waveform in each of the time slots changes phase continuously.