MXPA95004936A - Desmodulacion pi / 4-dqpsk in diversi - Google Patents

Abstract

The present invention relates to a method for demodulating and decoding signals modulated in the quaternary phase, comprising the steps of: receiving the signals modulated in the quaternary phase and producing a hard limited intermediate frequency signal, directly digitizing the intermediate frequency signal in phase hard limit, in order to produce a current of numeric values that represent instantaneous phase angles, compute the phase differences between pairs of numerical values separated by a quaternary symbol period, convert the phase differences using a sine / cosine look-up table produce pairs of values that represent pairs of soft data bits, and process the pairs of values using an error correction algorithm to give access to the credibility of the value pairs

Description

PI / 4-DQPSK DEMODULATION IN DIVERSITY BACKGROUND OF THE INVENTION La. present invention relates to radiocommunication systems in general and, more specifically to methods and systems for demodulating signals received in them if they are subjects. Radio-cellular systems are in the process of evolving from analog single-channel-frequency-carrier-frequency modulation (FDMA) systems that use continuous transmissions to multiple time-division access (FDMA) systems employing pulsed transmissions. In TDMA systems, communication is made between a base station and several mobile stations by assigning a single time segment in a time cycle to each connection. The discontinuous or packetized data transfer is uniformized to produce an understandable speech by using, for example, a buffer and a digital voice coding. The digitally coded voice is transmitted over the radio link using a digital modulation scheme such as Pi / 4-DQPSK. The use of Pi / 4-DQPSK for digital TDMA cellular communications is described in the IS54 standard of the Telecommunications Industry Association (Telesommunisa ions Industry Assosiation) of the United States of America.

Pi / 4-DQSPK is a modulation that transmits pairs of data bits by increasing the phase rotation of the signal vector by one of the following four angles: +/- 45 and +/- 135 degrees. During phase rotation, the signals modulated by Pi / 4-DQSPK show a frequency change equal to the phase change rate and this change in frequency can be measured to demodulate the data bit pairs. The modulation suffers a distortion in the transmission path due to the delayed echoes of the buildings, ets, which is why, in some cases, an equalizer is required in the mobile station. In cases in which the delay of that is limited, the equalizer can be replaced by a frequency dissolver. Frequency imiders are generally constructed using unsuitable analog components for integration into small silicon chips. One way, conventional systems combat such distortion is through diversity reception or transmission. In TDMA systems a variety of antennas have been implemented using either a diversity in the selection of antennas where antennas are tested to determine which of them provides the highest signal strength while the transmitter broadcasts data directed to another receiver to select the best antenna for the reception of the focused data, or the diversity combination where the signals received from the two antennas are added either before (pre-detession combination) or after the ion demodulation (post-detest combination). However, these diversity ethnic standards are problematic because rapid fading can change the determination of which antenna receives the best signal after having performed the test and selected an antenna. SUMMARY OF THE INVENTION According to the present invention, this and other inconveniences and difficulties are overcome by means of an exclusive digital method of demodulation ion of the Pi / 4-DQSPK signals using an analogue information converter in digital phase diresta are an algorithm of resepsión in diversity. Exemplary embodiments of the present invention also provide data quality measurements of data bits for the subsequent correction of error correction so as to enhance the symbiosis in diversity of two sanal receivers. Sonformity are exemplary modalities, a carrier wave is modulated by digital data using Pi / 4-DQSPK to transfer two bits of data by the sa bio of the radio carrier phase from the value at the end of the last symbol by a angle of + Z-45 degrees or +/- 135 degrees, these four possibilities represent the pairs of bits 00, 01, 11 or 10. The transitions of the signal transmitted by radio are filtered in the complex plane (I, Q) to limit the spectrum. In the reseptor, the resibida signal is reduced, filtered and amplified using a hard limiting medium (IF) intermediate amplifisator. The IF amplifier also produces an approximately logarithmic indication of the signal strength before limiting. The hard limit IF signal which is the phase information is fed to an analog to digital direct-to-digital converter. A numerical value representing an instantaneous phase is produced and processed to demodulate data symbols. The demodulated data symbols may be combined with the logarithmic signal intensity information for supplying data bits with quality information to an error corrector. In accordance with some of the exemplary embodiments of the present invention, two antennas, receivers, IF amplifiers and analog information converters are provided in digital phase. The numerical phase values and the logarithmic values of the signal strength are successively reproduced to produce demodulated data of improved quality which is fed to an error correction decoder. The dessodifisads data may represent a digitized signal and are then fed, for example, to a signal senal dessodi to produce a local analog signal to drive a telephone handset or speakers. BRIEF DESCRIPTION OF THE DRAWINGS The foregoing and other objects, features and advantages of the present invention will be more readily understood upon reading the following detailed symmetry in symbology with the drawings in which: Figure 1 is a block diagram of a Pi receiver. / 4-DQSPK of a single sonality sanal with an exemplary embodiment of the present invention; Figure 2 is a block diagram of an exemplary dual diversity exemplary diversity are the present invention; and Figure 3 is a block diagram of an exemplary configuration of the signal processing unit of Figure 2. DETAILED DESCRIPTION Figure 1 shows a monosanal receiver array of sonicity are an exemplary embodiment of the present invention. A sonvepsional method to process the signals transmitted by radio of arrival includes the resepsión, amplifies ifas ion and separation of the signal in Cartesian components in phase and in quadrature (I, Q) and its sonversion of analogical information in digital to provide a flow of complex numbers for processing. An advantageous alternative to the conversion of analog information into digital Cartesian is disclosed in US Pat. No. 5,048,029 to Paul W. Dent, entitled "Logpolar Signal Processing", said patent is hereby incorporated by reference. In the logpolar method, the signal is converted from digital to digital information, not Cartesian components (I, Q), but as polar somponents that include a phase angle and a logarithmic measurement of the amplitude of the vector or radius. If desired, these values can be converted numerically into Cartesian components (I, Q) after their digitization using appropriate digital signal processing. An advantage of the logpolar digi tal i zassion is the achievement of a high dynamic range of instantaneous receiver without the use of automatic gain control. US Patent No. 5,034,669 to Paul. Dent, entitled "Direst Phase / Frequensy Digitizing" (Digi such phase / Diresta Frequency Digit) dessribe different methods to directly digitize the phase of a signal transmitted by radio are hard limitation, said presentation is also incorporated here by reference. US Pat. No. 5,220,275 to Peter Holmquist entitled "Assumption Phase Digitizer" (Digi tal of Phase of Asurauiadsr) discloses another method of digi tal i zation diresta phase that can be used to implement the present invention and is also incorporated herein by reference . U.S. Patent No. 5,335,250, entitled "Bidirestional Demoduiation Method and Apparatus" (Method and Desire Device, Bidi ressional) dessribe the radio demodulation of radio TDMA packets by means of the digi tal izasion of the signal, its recording in a memory and then its demodulation direstates or inverse from an initial point of known data symbols surrounding dehonose data and the separation of results of the demodulation address that provides data of the highest quality. In the present, the direct phase measurement is defined to include any device or method to produce numerical values of the instantaneous phase of a complex signal that does not first insulate the separation of the signal into real and imaginary components or Cartesians and the satire. of an arc tangent. In Figure 1, a signal is received from the antenna 10, which is reduced, filtered and amplified in the superheterodyne receiver 11 to produce an instantaneous signal of an instantaneous signal strength (RSSI) 12 and an intermediate frequency signal of hard limitation 13 to the logpolar digi tal hoist 14. The logpolar hoist digi 14 can operate in accordance with the aforementioned US Pat. No. 5,048,029 and its suffix minus a direct-phase digital hoist 15 such as is disclosed in the documents before mensioned that are insorporated by reference. Of sonformity are a first exemplary embodiment of the invention, only the phase is digitalized and Pi / 4-QPSK is demodulated. In another exemplary embodiment that will be discussed later, the RSSI signal is also digitalized in the converter from A to D 16 and both the phase and the RSSI signals are processed together to supply the demodulated bits. informacion de salidad to a dessodi fisador of correction of errors. The digi talizer of phase 15 produces a flow of numerical values that represent the instantaneous phase of the signal in the form of modulo-2Pi. The modulo-2Pi form indicates that the entire numerical value K representing the phase angle can take values between 0 and M-l to represent the phase angle K / M x 2Pi. An integer value of K = M would therefore represent the angle 2Pi which is the same as the angle 0. In the arithmetic of modulo-M, the value M and 0 are represented somo corresponding identical to the identity of the angles 0 and 2Pi. More specifically, using the binary arithmetic, for example, of an 8-bit word length, the values 00000000 (0) to 11111111 (255) represent angles from 0 to 255/256 x 2Pi. When you click one. unit to 11111111 we obtain (1) 00000000 where the digit 1 more to the left leaves the range of 8 bits and is omitted leaving the value OOOOOOOO (0) that sorresponde to 0 degrees again. Therefore, the circular operation of arithmetic of finite words is represented in the symmetric phase angle domain. The Pi / 4-DQSPK signals are designed for sonority in this exemplary mode by using a phase-module diffuser-2Pi 17 that subtracts the digital phase value from a symbol time before the astual phase value to obtain a value of phase difference. The different phase value is used as an address for a look-up table 18 ssnestada to the input of digital logic, its output is two bits of demodulated data that sorresponden to the suadrante of 90 degrees in the sual is the phase difference. It is desired that the demodulation not only produce bits of demodulated data but also an indiscrimination of its reliability to the error correction de-modifier 19. This can be achieved by using the look-up table 18 storing values corresponding to the cosine and sin of the difference of phase, respectively. These so-called soft values indicate a bit polarity of the quaternary symbol demodulated in the sign of the soseno component and the other bit in the sign of the sine component. Sine and sine magnitudes do not allow for the saliency and sredibility of the bits. The nominal phase differences between consecutive symbols using Pi / 4-DQSPK are +/- 45 or +/- 135, that is, they are in the middle of the quadrants. The magnitudes of the cosines and sines of such angles are 1 / root (2). If a phase difference is found around 0 degrees, the soseno will be + 1 which gives a reliable indication that its sign is positive, while the Sine magnitude is approximately 0, which indicates a very unreliable determination of the sine sign. These soundness measurements can be employed in the error correction probe 19 to help determine and correct probably erroneous bits. The error correction de-modifier 19 processes these symbols representing the sodified information using disambiguation measurements to produce deodorized symbols. Several symbols that represent less accurate information can be transmitted without being sodifiscated and therefore the dessodi fisasor of correction of errors does not dessodifisa. In the last exemplary exemplary embodiment, the RSSI signal is also digitalized and taken into consideration to determine the number of bits. The RSSI signal can be combined after its digi tal izasión are the phase value and be converted in Cartesian Z = (I, Q). Representing the value for a symbol period by Z (i-l) and the astual value by Zi, the value Zd = (Xd, Yd) = Zi is formed. (Zi-1) * where * denotes complete play. The real part of Zd (Xd >; then represents the "soft" value of a data bit and the imaginary part Yd represents the soft value of the second data bit of the quaternary symbol number "i". Alternatively, in polar representation, Xd and Yd can be expressed from the following f rma; Xd = Ri.R (i-l). C0S (Ai-A (il)> Yd = Ri.R (il) SIN (Ai-A (il)? Where Ai represents the phase values and Ri represents the corresponding amplitude values. represented by the symbol "." are converted into adiations using logarithmic values, therefore: LOG (XD) = LOG (Ri) + L0G (R (i-1)) + L0GC0S (Ai-A (i-1) ) LOG (YD) = L0G (Ri) + L0GiR (i-1)) + L0GSIN (Ai-A (i-1)) Since the RSSI value occurs in logarithmic form corresponding to the LOG (R) values and the Query table 18 contains values of L0GC03 and LOGSIN of the angle differences, instead of COS and SIN, the soft-bit values array can be made are the previous equations using only short words, fixed-point add-ons and without multipliers (not it shows).

In this case, the logarithmic values of Xd and Yd can be converted into linear values by means of an anti log table, but it may be convenient to retain the values in logarithmic form in order to represent a wide dynamic range of values with long words. The subsequent error correction device that processes the soft values can easily be arranged to be logarithmically accepted and used either in a log-arithmetic unit or converted into linear values by means of an antilog function. To obtain good performance it is necessary to sample and digitalize the arrival signal at the optimal points within the symbol period. These sampling times must correspond to the phase angle that has just completed its rotation in +/- 45 or +/- 135 from an old value to a new value. Optimum sampling instants can be determined by sampling in, for example, uniformly spaced points in the symbol period, and then deciding which of the sampling phases is used to demodulate the sonicity are the values to be evaluated. A signal segment TDMA comprises, for example, a first pattern of symbols sonosidss to the pripsipio, a number of soundless symbols to be determined, and a second pattern of symbols sonosidos at the end. The first pattern and the second pattern of sonosidos symbols are known to veses somo "words of sinsronismo" (synswords). By means of the sorrelation of the osho possible phases of sample of the received signal are the first sonoside pattern and selecting the sample syndromism that provides the highest sorrelation, the optimum sample sinsronisms for the demodulation is determined. For example, representing samples taken are a difference of 1/8 symbol by means of Zl, Z2, Z3 Zi and the sample values that the sound symbols should produce are TI, T2, T3 Ti, the values of sorrelation are calculated Cl = Zi.Tl * + Z9.T2 * + Z17.T3 * -I- Z (? N-7) .Tn * C2 = Z2.T1 * + Z10.T2 * + Z18.T3 * + Z (8n-6) .Tn * C3 = Z3.T1 * + Z11.T2 * + Z19.T3 * + Z (8n-5) .Tn * The complete correlation having the greatest magnitude determines the sampling phase selected for demodulation. We assume for example that C3 has the greatest magnitude, then the subset of samples Z3, Zll, Z19, Z27 etc. is essoge. for demodulation. Due to the sample interval of 1/8 of symbol, it is observed that the subsessive values of Z employees now have indices separated by osho instead of one. Then, the soft-bit values are calculated using: Zd = (Xd, Yd) = Zi.Z * (i-8) or, in polar representation: Xd = Ri.R (i-8) .COSÍ Ai-A (i-8)) Yd = Ri.RÍ i-8). SIN (Ai-A (i-8)) or, in logarithmic form: LOG ( XD) = LOG (Ri) + LOGY (i-8> + L0GC0S (Ai-A (i-8)) LOG (YD) = L0G (Ri) + L0GiR (i-8)) + LOGSIN (Ai-A (i-8)) When it is taught lofts in the sanal, it is. it is possible that the first word of sinsronism present alterations preventing the determination of the best phase of sampling. Similar to the dessrita in the North American Patent No. 5,335,250 above insorsed, the values of sorrelation can also be salsulated using the second word of sinsronism. According to the relative sredibi lities of the correlation results, the data can be demodulated in direstión diresta, es desir, from the first word of sinsronismo ssmo referensia, or in an inverse diressión from the second word of sinsrsnismo, or in both directions. Before defining a measurement of the sredibility of a word of sinsronism, a method to determine how such values can be used to choose one or more demodulation directions will be presented below. It is assumed, for example, that the values of word sredibility of sinsronism are defined in some way and are quantified in a limited number of cases such as high, medium and low probability. Then the combination of the values of sredibilidad of the two words of sinsronismo provide nine different sasos. For each saso, the optimal demodulation ion strategy that provides the least number of symbol errors on average can be determined in advance by simulation and a destionation table containing nine entries can be integrated into the team. For example, the table of desision in this simplified sample would be: T A B L A 1 Word Credibility of Sinsronism Number 1 Low mid High Credi- Ba- Semi-diresta, Diresta Diresta 1 ith ja Semi-reverse Word Me- Inverse Semi-diresta Diresta de dia Semi-reverse Sinsro-nismo Al- Inverse Reverse Semi-diresta 2 ta Semi-Reverse To determine which is the optimal demodulation direction to fill the previous table, the following procedure can be performed. A computerized simulation of the transmitter, fade sanal and receiver is performed and the received TDMA signal segments are classified by ssnformity are the nine possible instances of credibility of synchronism word previously defined. Then, for each category, the number of symbol errors is separated out separately for the three steps of the ion-demodulation, ie desir, diresta, inverse or semi-di straight, serai-inverse, abbreviating now, three-phase sasos are the letters F , B and H. The number of symbol errors found are assumed separately in 3 times 9 or 27 counters that are related to the sade of demodulation used. They are sada of pairs of saliency pairs of sinsronism. For each category, the desraodulasion direction F, B or H that gives the smallest number of symbol errors for this category is then selected to be placed in the previous desideration table. As an additional development of this novel process, it will now be explained how the word reliability of sinsronism can be optimally detected in a limited number of levels, for example, the three "low", "medium" and "high" sredibility values previously postulated .

We suppose that we have a means ion not quantified or continuous of the credibility. This can, first, finally be quantified in a larger number of sasos, for example 16. Therefore, there are 16x16 = 256 possible combinations of the sredibility of the two words of sipsronism. Extending the simulation prosßdimißnto to salsular the total number of symbol errors in each of the 256 categories of signal output using sada. Demodulation resolution in 3x256 counters, and then determining for sadategy what demodulation strategy F, B or H provides the smallest number of errors, a table similar to table 2 is created that is presented at sonu inuasión.

TABLE 2 O 1 2 3 4 5 6 7 8 9 absdef 0 HHFFFFFFFFFFFFFF 1 HHHFFFFFFFFFFFFF 2 BHHHFFFFFFFFFFFF 3 BBHHHFFFFFFFFFFF 4 BBBHHHFFFFFFFFFF 5 BBBBHHHHHFFFFFFF 6 BBBBBHHHHHFFFFFF 7 BBBBBHHHHHFFFFFF 8 BBBBBHHHHHHHFFFF 9 BBBBBBHHHHHHHFFF to BBBBBBBBHHHHHHHF b BBBBBBBBHHHHHHHF s BBBBBBBBBHHHHHHH d BBBBBBBBBBHHHHHH and BBBBBBBBBBHHHHHH f BBBBBBBBBBBBHHHH In practice, variations The statisticians can not provide a perfectly symmetric table as in the previous table, but due to theoretical reasons to wait for a time reversal symmetry, one can first correct a non-symmetric table to obtain a symmetric table. For example, if the entry in row 2, solves 1, would have been B instead of H, the error penalty penalty can be examined by entering the row 2, sounds 1 in H compared to the input of the entry in row 1, set 2 in F, and choose the correction with the minor izasióp penalty. For example, we suppose that the simulations for the category of saltiness of sinsronism of row 2, solumn 1 gave the following results: TOTAL COUNTING OF ERRORS DIRECTION OF REMOVAL 4819 F 3116 H 3102 B In this case, the selection of the error plus low would have provided B in the row 2 position, column 1 that is not symmetrically compatible with an H in the. row 1, solumn 2. However it would have been sompatible are an F in row 1 solumn 2. We now assume that the simulations for row 1, column 2 have provided the following: TOTAL COUNTING OF ERRORS DIRECTION OF DEMODULATION 3018 F 3009 H 3917 B The criminalization in recession is the change from H to F in row 1, it assumes 2 to reestablish the symmetry would have inscribed the total of errors from 3009 to 3018, a penalty of 9 additional errors. Alternatively, we could have restored the symmetry by selecting H in row 2, solumn 1. This would have inscribed the total number of errors from 3102 to 3116, an inset of 14. Since this last solusion is worse, the previous form pair restore symmetry. Then, the table can be divided into approximate regions in which the demodulatory strategy is approximately uniform, as follows: T A B L A 3 0 1 2 3 4 5 6 7 8 9 a b s d f f F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F F 6 B B B B B H H H H H F F F F F F 7 B B B B B H H H H F F F F F 8 B B B B B H H H H H H H F F F F 9 B B B B B B H H H H H H H F F F B B B B B B B B H H H H H H H F B B B B B B B B H H H H H H H H F s B B B B B B B H H H H H H H H H d B B B B B B B B B B H H H H H H e B B B B B B B B B B H H H H H H f B B B B B B B B B B B B H H H H The strategy desmsdulasión in sada region Table 3 after sambia to the predominant strategy giving, for the nine regions: H F F B H H B B B as originally assumed in table 1. As illustrated above, the penalty of number of errors can be determined by changing an entry F, B or H in table 3 in the dominant entry for one of the nine regions. Therefore, the total penalty can be saved to establish the same type of demodulatory F, B or H to be used consistently in each region. This sort of penal penalty can also be repeated for every possible position of the two horizontal or vertical dividing lines in Table 3 that delineate the nine regions. The positions of the four dividing lines are then calculated, which provide a lesser penalty in the number of errors in order to establish soundness in one of the nine regions. Therefore, the optimum supersession of the synchronism output values 0, 1, 2 e, f has been systematically determined in only three wider regions that can be called "GOOD" "AVERAGE" and "BAD", the result it can be permanently integrated into the equipment logic to carry out the optimal quantification, thus they are the decision of the type of demodulator (F, B or H) that was optimal for one of the suantifidated regions. The limits between the nine regions have been determined in a systematic way by evaluating the increase in the total number of symbol errors for each position of the possible limit and to unify the demodulation strategy in each region in F, B or H. The positions of The limit and the best demarcation strategy for each region is therefore determined systematically to obtain the minimum net number of symbol errors. This explains how a word sredibility edict of continuous syncronism can be roughly sutured to reduce the size of the query table required to have ion demodulation strategies. Those skilled in the art will note that tables of musho size greater than 3x3 entries can be used in the prastics are the modern memory of the memories. The above example employs tables of size 3x3 only for the purpose of illustrating the present invention but the procedure may be employed are larger tables such as 8x8. A sontinuasión dessribirá sóms can be obtained a measure of credibility or outgoing word of sinsronismo. When a differencial phase demodulation is used above, the purpose of the correlation with the word of sinshonism is to define the best synchronicity of sampling are the symbol period. The best phase of sampling is the phase that would result in the least number of errors, it is desir the silliness that provides the greater relasión between signal and (noise + iterferrensia + ISI). If the word readers of Sysronism Ti employed for the sorrelation are of unit length, then the correlation is the highest magnitude indicating the best sampling phase. Likewise, the square of the indistinct magnitude how the energy of the resibed signal correlates with the word of sinsronism while the difference between the total energy received and the word energy of syncronism represents all the unwanted energy. If, for example, C2 represents the major correlation, then the total signal energy in this synchrony is: E = Z2.Z2 * + Z10.Z10 * + Z18.Z18 * + Z (8n-6) * .Z (8n -6) The desired content of signal energy in this signal is S = | C2. Therefore, a medicion of the saliency of the word of sinsronism is S / (E-S). It is a numerical numerical value that can then suant i fuse of sonformity are the method presented above. It is evident that alternative measures of the word saltiness of sinsronism can be defined. For example, once the optimum sampling phase is determined, this can be used to demodulate the symbols of the word of sinsronism; If the results are the values given a priori, the number of symbol errors in the synchronism word can be determined and a scission measurement of the word of sinsronism can be used. It has been shown how measurements of the saliency of the synre- dlation of syncronism can be obtained, how saliency can be approximated in a limited number of values, and how the approximate salience of two synchronized words can be used to deduce the Optimum ion demodulation strategy for dehonose data located between them. It is observed that the described procedure can be applied to decide between any number of alternative strategies of ion demoditions, and not only between the diffraction phase diffraction diresta or inverse are different phases of sampling. For example, another demodulation strategy that could be included in the available repertoire can be a suasisoherent algorithm. In suasicoherent demodulation, a desssons data symbol is demodulated not by the use of sampling from a previous symbol period but by using an estimator of the signal lestor for a previous symbol period. For the first symbol of deshonosido data (is desir, the closest to the word pattern of sinshonismo sonosido) this estimate is derived from the correlation result chosen 'Cj' by calculation: Zref = CjTn / n. The first data symbol is then demodulated by means of: Zd = Z (j + 8n). Zref * Zref is then astualized by rotating at an angle of +/- 45 or -I - / - 135 degrees as indicated by the suadrant where Zd is found, and then by adducing a frassion of the difference of Zref and Z (j + 8n). For example, we assume that Zd is in the first suadrant. Then it is calculated: Zref = Zref. EXP (jPi / 4) Zref = Zref + (Z (j + 8n) -Zref) / 16 to obtain a new value of Zref to demodulate the second data symbol. Because the Zref vector is carried either directly or inversely according to the direction of the stratified demodulators, the suassociated scheme shows more dependence of the direction than simply the separation of the syncronized sampling ion from the first or second word of synchronism. In the scheme suasisohßrent ?, different numbers of symbol errors will be obtained for the two directions even though the sampling phase given by the two words of sinsronisms is the same, whereas in this case the demodulation of the differential phase is dependent on the diress . Therefore, it can be expected that the selection of the ion-desorption stress given by the pre-calculated decision table is more significant. As an adisisnal demodulation strategy, an echo equalization algorithm can be provided. When a correlation shows a high energy level at + or - a symbol period of the correlation peak, this indicates echoes or time dispersion that may be required to compensate for a correct demodulation. The dessripsión of all the possible sonorussisnes of equalizers is beyond the alsanse of the present invention, nevertheless, they are the purpose of giving a somatic information will dessribirá a step that can be taken to reduce the effect of echoes. It can be shown that the effect of echo or time dispersion is, among other bland ones, of adding a somple constant to the result of the saule: Zd = Zi, Z (i.8) * in such a way that the possible values of Zd are no longer distributed around a circle circle (0,0) but around a displaced center. The substraction of this Zd's skeletal constant before determining the quadrant where it is located or before passing its real and imaginary parts as flat bit values to the error correction de-modifier results in a reduction of symbol errors. The choice to go or not to this strategy may depend on the correlation results from which an estimate of the correct displacement can be made. Comparing this estimated displacement with a threshold value, it can be approximated as "signi fi cant" or "insignificant" in an ana- lystically determined way for the previously dessritated su- nity of the salinity of sanshonism. This sodifission is then used to determine whether the displacement sorption is used. In general, the method allows for the cessation of the TDMA packages supported in the 36 possible cases of High, Low and Medium out of sync for both words of multiplexed synronism by the two passwords per word of synchronization of significant displacement. / insignificant due to time dispersion, and the on-line evaluation of all the available demodulation strategies for sada slasi fisasión are the object of being able to build a table of students that indicates the best average strategy to be used for sada saso, the table salsuiada is then integrated into the team. The exemplary embodiments described above may be extended to cases where more than two known symbol patterns are included, for example, by use of what is known as the Digital Voice Color Code (DVCC) defined in the TIA IS54 standard. a symbol pattern is sounded as are the words of synchronism. Such a case can be treated as two sasos of data dessonosidss joined by patterns of data conosidos, and the present invention can be aplised separately in sada saeo. The above-described exemplary embodiment may also extend the diversity reception where the choices for demodulating unknown data include the selection of the signal source between several antennas. Figure 2 shows a dual diversity diversity receiver is another exemplary embodiment of the present invention. The antennas 20a and 20b nominally receive the same signal but due to slight differences in position or polarization between them, the ratios between signal and interference are not identical. The two signals are amplified, filtered and reduced in superheterodyne 21a and 21b receivers, before being able to be digitized by converters from A to D 22. The digital signal values passed to the data processing system processing unit. these values of sonformity are what is described at sont inuasión. It can be sampled from sentences of syncronism in the same way as it was previously dessibed for the unisex sanal, sass in non-diversity, but this time for signals from both receivers. The correlation qualities of sinsrsniems are calculated using, for example, one of the methods described above, the synchronization qualities roughly categorized by comparison with predetermined thresholds, and the combination of optimized synchronization qualities for all the correlations of sinsronism employed for Selession a demodulation ion strategy and the separation of a signal source to demodulate. For example, if the TDMA signal segment consists of two symbol patterns (symbols of non-symbolism), they are symbols of unsophisticated data between the loe, then the selection of di-modulation, inverse, or semi-straight, is reversed by the The use of "a" or "b" antenna signals provides for possible choices. The best efficiency will be based on suing menoe lae suatro synchronism quality measurementsAs described above, the function that dessribe as the best resolution of ion desmsdulas varies depending on the qualities of sinsronism can be pre-salsulated by off-line simulation using a demodulation strategy to demodulate TDMA signal segments of all kinds of salinity. Unsympathetic, and the strategy that provides the lowest number of symbol errors for the class can then be recorded in a team table that can be integrated into the team. For example, there are still only two antennas, the number of signal source choices can be artificially increased by forming, within the digital signal processing unit 23, a weighted euma or difference of the signals of the two antennas. The sum and the difference can be treated as somersensors and assorted speakers of signal sources, or as the two primary candidates.

Likewise, the weighting factors used to calculate the sum or difference may depend on the correlations calculated for the two original signals. Especially, the correlations provide information on the relative phase of the signals of the two antennas, which can be used to determine weights in such a way that a positive admission can be carried out. Also, the weights that can be derived from correlations are words of sinsroniemo that are found before or after the unknown symbols of data. Accordingly, a weighted sum formed of a first set of weights and a weighted sum formed of a second set of weights can be considered as two alternative sources of signal as well as the two original signals. Each of these smooth signals can be chosen to demodulate directly from the first word of synchronism, in reverse form from the second word of sinsrsnism or half and half. This provides a total of eight possible ways to obtain demodulation results from the first half of the unknown symbols and eight possible ways of obtaining demodulation results from the second half of the unknown symbols. The choice of the word of unscripturalism located further to the left as a reference of deemodulation can be excluded when the words of sinsroniemo are used more to the right to form the weighted admission, and vice versa. However, it is not necessary to exclude this possibility from offline simulation; To the sontrario, the results of the simulation off line will verify this soneiderasión or no. As before, the elimination of the desraodulasion strategy is pre-saved in function of the output of approximately six-digit sentences and integrated in a digital or digital table that is part of the digital signal processing unit. Figure 3 shows a block diagram of an exemplary signal processing unit 23 for the reseptor of Figure 2. The digital signals of the two antennas are first recorded in a buffer 30 from which they can be resuperadae by a sorrelasiopador of einsropiemo 31. The self-replicator of sinsronismo resuperates the samples of the first and second sources of signal for its sorrelasión with the first word of sinsronismo and later recovers samples of the first and second sources of signal for its correlation are the second word of sinsronismo produsiendo asi suatro sets of results of sorrelasión. Each set of correlation results includes less an indiscrimination of the sample in s that gives the maximum correlation, the complex value of the maximum correlation, and a correlation salinity measurement, but may also include the correlated value of the complex correlation in all sample phases. The results of the sorrelation are sent to the strategy selection unit 32 and to the weighted addressee 33. The strategy separation unit also provides the possibility of several possible algorithms or demodulation strategies that can be performed by the demodulator 34. The data symbols demodulated with the soft-sound information from the demodulator 34 to the error corrector deesodi 35 which can perform interleaving operations to disperse the errors over time. The weighted adder 33 can operate in one of two general ways depending on whether the samples of the two signal sources are a time offset between the lae are ssmbined or not. This selection can either be predetermined or it can be function of the qualities of eipsirsniemo already propsrs ionadae to the unit of selection of eetrategia 32. For exemplary modalities where the combination of samples of the two signal sources occur only for samples taken at the same time , the weighted sum is calculated as follows. For the sample fae "i", which provides the correlation Ca (i) in one source and Cb (i) in the other source, the weighted sum is formed as follows: Rel (i + 8k) = Ra (i + 8k). Ca (i) * + Rb (i + 8k). Cb (i> * for i locating above the sample faee number (for example 1 to 8) and k being located above the number of symbol periods in the memorized signal segment. symbol period (for example, 8 samples) inisiando are the sample number 'i' are combined using sorrelation values 'i' as weights.The weighted sum Rsl returns to memory where it can be recovered by means of the correlation calculation 31 and a The set of correlation results are calculated in the same way as for the original signals and are supplied to the strategy separation unit.This is also done by using the results of sorrrelation for the second word of syncronism to obtain Rs2. in the sual ee allows a combination of lae doe signal sources with a time offset, the best sample phase 'i' that provides the maximum correlation are the first source The signal phase 'j' of the second signal source is as follows: RslOk) = Ra (i + 8k). Ca (i) * + Rb (j + 8k). Cb (j) * for all k. In this case, only one sample fae for the sum is produced in such a way that no additional sorrelation is required to determine the best sample phase. An Rs2 network is calculated in the same way using a Ca and Cb calculated based on the second word of sinsronism. The 'i' phase of Ra, the 'j' phase of Rb and Rsl and Rs2 are the available candidates for direst or inverse demodulation. In this case, the signal loss and demodulation strategy is predetermined only by the results of the original smooth correlations, so that the weighted sum Rsl or Rs2 requires a calyx only if the indexed strategy needs them. As a third way to determine the weighted sums, it is possible to combine in diversity the two signal sources Pi / 4-DQSPK after their differential demodulation but before the detection of symbols as follows: Rs (8k) = Rat l + 8k ) .Raí i + 8k-8) * + Rb (j + 8k). b (j + 8k-8) * This has the desirable property that the contribution of each term falls as the square of the signal amplitude. Likewise, the construstive addition occurs independently of the signal phases received. Therefore it is also possible to effect the combination in diversity using this method before effecting synchronization sorrelation. This can be achieved by the following pattern: Rs (i) = Rad). Raí i-8 > * + Rb (i). Rb (i-8> * The synchronization symmetry is then effected in Rs to determine the symbol's nsroni zation In this case, because Rs is a network of different signal samples, the correlation is made by using patterns of the known differential symbol, and not by the signal vectors sounded in absolute form, as above.The words of differensial sinsronism are provided by Ti.Tíi-D * and are a symbol more sorta than the word of absolute sinsronism. with a first and second sonoside differential symbol pattern can be effected, quantized output measurements can be generated and used in the selection of address or demodulation mode.A sonicity receiver are the exemplary embodiments of the present invention can provide several demodulation modes without antenna diversity An exemplary block diagram for such a receiver would be the same as in Figure 3, but only you with one entry, Ra or Rb, to the buffer 30 instead of two. The exemplary embodiments described above have the purpose of illustrating the present invention in all its aspects but not of restricting it. Therefore, the present invention may undergo several operations according to the details of its implementation that a person skilled in the art can derive from the description contained herein.

All of these variations and modi fi ed ions are considered within the scope and spirit of the present invention of sonformity with what is defined in the following re visions. NOVELTY OF THE INVENTION Having dessrito the present invention, a novelty is soneid and, therefore, resides in property contained in the following lae

Claims (47)

1. CLAIMS 1. A method to demodulate and de-modulate signals modulated in the suaternary phase that consists of the steps of: receiving dishas signals modulated in the suaternary phase and producing an intermediate frequency signal of hard limitation, digitalizing direstamept? in phase disha signal of fresuensia intermediate of hard limitation to produce a flow of numeral values that represent instantaneous angles of phase, salsular the differencies of fae between pairs of disks numerical values spasiados a period of suaternary symbol between elloe, sonvertir dishae diferennsi e de faee using a table of questions. sine / cosine to produce pairs of values that represent pairs of data bits, and process dishss pairs of values using an error correction algorithm to produce data.
2. 2. An apparatus for demodulating and desmodifying signals modulated in a quaternary fare comprising: means for receiving signals and producing an intermediate signal of hard limitation, medium of digi tals of the output line to convert disha signal of intermediate frequency limitation lasts in a flow of numerical values that represent instantaneous phase angles, a phase subtractor to determine phase differences between pairs of dish numerical values with a given interval of time, conversion means to convert those phase differences into pairs of values which represent pairs of data bits, and error correction means for processing even pairs of values to produce data.
3. 3. A sonicity method is the resonance ion 1 in which said modulation in the suaternary phase is QPSK difensial Pi / 4.
4. 4. A method of compliance is claim 2 wherein said quaternary phase modulation is differential QPSK Pi / 4.
5. 5. A method for transmitting information between a first station and a second station on a radio station system comprising the steps of: assembling pairs of bits of sodifised information of error error in symbols alternating ios, changing a unit signal of unitary amplitude from an astual position to a new position rotated by 45, -45, 135 or -135 degrees in relation to the asthe disha sion of conformity are a value of said suaternary symbol, apply real and imaginary components of a sequence of dish vectors of complex signal to low pass filters to obtain continuous waves I and Q, modulate the carrier waves of cosine radius and sine are dishas I and Q waves and transmit a sum of modulated carrier dishas somo a signal sompleja from disha first estaión, reshaping disha signal in disha second somatic stain and producing an intermediate frequency signal of hard limitation, prssesßr disha intermediate fresuensia signal using a digester phase digitor to produce a flow of numerical values that represent instantaneous angles of signal phase, calculate the phase differences between pairs of dish numerical values spaced by a quaternary symbol between the, dishas phase differences in pairs of values representing said pairs of information bits, and process discrete pairs of information values using an error corrector dessodi to obtain data.
6. 6. A method of conformance is claim 5 wherein the conversion step further includes the step of using a sonar table.
7. 7. A sonicity method is the reimagining 6, where said table contains the sine / cosine values.
8. 8. A method of sonification is claim 5 wherein said application step further comprises the step of supplying low pass filters having a response of fresuensia which is a suquired root of Nyquist filter for a given number of symbols.
9. 9. A sonicity method is the re vindisas ion 8 where the Nyquist filter has an increased cosine response.
10. 10. A method for de-modulating signals modulated in the quaternary phase comprising the steps of: receiving said signals modulated in the quaternary phase and producing a signal of intermediate freshness of hard limitation and an intensity signal that indicates approximately the intensity of the signal resibidae, disvert signal of hard limitation and disha signal of inteneity in a flow of numbers somplejoe using a logipolar digi talizer, salsular a produsto somplejo of pairs of dishos somplejos numbers spaced by a period of symbol suaternario among them are sonjugation of one of each pair of complex numbers to produce different spatial numbers, and process different complex differential numbers using an error correction algorithm to produce data.
11. 11. An apparatus for the demodulation of modulated signals in the quaternary phase which comprises: reception means for receiving the modulated signals and producing a signal of intermediate freshness of hard limitation and one. Indication of approximately log signal intensity, means for digitizing said intermediate frequency signal and disha indication of signal intensity in a stream of complex numbers, means for displaying a produse of pairs of dishoe numbers, sparse numbers spaced by a Swarth symbol among them for to produce differentiated differential numbers, means of dessodi fi cation of error errors to prosecute different serial numbers to reduce the number of transmission errors.
12. 12. A method of sonification is the claim 10 wherein disha msdulasión phase suaternaria ee QPSK differencial Pi / 4.
13. 13. A sonicity device is the claim 11 where disha modulation in suaternaria phase is QPSK differencial Pi / 4.
14. 14. A method for transmitting information between a first station and a second station on a radio somenication system comprising the steps of: assembling pairs of coded information bits of error correction in quaternary symbols, switching a unit signal vector to Starting from an astual position to a new position rotated by 45, -45, 135 or -135 degrees from disha s astual position of sonformity with a value of disho symbol suaternario, apply real and imaginary components to a sequence of said signal vectors With low pass filters to obtain continuous I and Q waves, modulate carrier waves of Fresenius radio cosine and sine are dishas I and Q waves and transmit a sum of modulated carrier dishas somo a signal sompleja from disha first estaion, reshave disha signal in the second phase and produce an intermediate frequency of hard limitation and an approximate indication ximately logarithmic signal intensity, log logically converts the signal of fresuensia intermediate of hard limitation and disha indesection of signal intensity in a flow of numbers pleple, salsulate a produsto somplex of pairs of dish numbers, sparse numbers spaced for a period of quaternary symbol between the loe, one of which is conjugated to produce di fferential di fferential numbers, and to proceed with di fferent di erential numbers using an error correction function to obtain data, 15. A method for transmitting information between a first, station and a second station in a radio communication system comprising the steps of: assembling pairs of bits of sodified information of correction of errors in quaternary symbols, interleaving said quaternary symbols between a first and second symbol patterns sonosidos in a reseptor for . obtain groups of symbols for transmission, transmit said groups of symbols using Pi / 4-DQPSK signals, hold dishas signals transmitted in the second station and produce a flow of somple numbers from the la, correlate disho stream of full numbers are disho first and second patterns of eimboloe to determine a first and second if nsroni zas ion that provides the. Maximum correlation are first and second symbol patterns, respectively, and a first and second output value of sorrelation sounder, prosearse first and second values of sorrelation output to determine whether first or second synchronization is a preferred synchronization, the second produsto complex of pairs of dish numbers somplÃos spaced by a period of quaternary symbol between them starting at a location in disho stream of somplÃos numbers indexed by said if nsroni zas preferred ion, being played a pair to produce differential complex numbers, and Process dishonomic somplitudinal numbers using a dessodi error corrector to obtain data. 16. One method of sonification is the re vindisas i ón 15 where said preferred synchronization is equal to disha first sansroni zasión for a first number of products sausulos and equal to disha second sinsroni zas ion for a second number of prodigal saules. 17. A method of soundness is the rei vi ndisasi ón 15 where one of these symbol patterns is a code Codified Digital Voice Color (CDVCC). 18. A method for transmitting information between a first station and a second station that includes: assembling pairs of bits of sodifised information of error sorption in quaternary information symbols, intercalating rooms and quaternaries between a first and second symbol patterns of the receiver is words of syncronism to obtain groups of symbols for transmission, transmitting groups of symbols using Pi / 4-DQPSK signals, receive said signals transmitted in said second station and produce a flow of sompleous numbers from there, correlating said flow of sompl numbers are said first and second words of sanshonism to determine a first and second if nsroni zas ions that provide a maximum correlation with disha first and second words of sinsronismo respectively and a first and second value of salor de sorrelasión corresponding, prosesar dishos first and second values of c correlation ality to determine a preferred desmsdulation algorithm among various design algorithms, demodulate dishss of sodifised information bits of errors using the preferred demodulation algorithm to produce demodulated information symbols, and prosecute dish demodulated infrastructures symbols using a dessodi fsador of sorressión of errors to obtain dishos bits of information. 19, A sonicity method is the re-vindisation 15 wherein the step of prosessing the first and second qualities of sorrelation further involves the steps of: quantifying approximately said first and second salinity values using predetermined thresholds to produce one. first and second indisasion of approximate saliency, combine dishas approximate quality indications to obtain a diressión in a. memory, and use disha sion to access memory to obtain a value indicative of disha if nsroni zac ion preferred. 20, A sopformity method is the re visions iop 18 where said first and second correlation salt processing step further includes the steps of: quantizing approximately first and second salinity values using predetermined thresholds to produce a first and second indication of approximate quality, combine dishases of approximate saliency to obtain a diression in the memory, and use dishair diressión to have access to disha memory to obtain a value indicative of disho algorithm of deemodulaci ion. 21. A conformity method is the re-vindisation 18 wherein disha plurality of algorithms available includes: d? Ethpodulation of fae difensial using a first or second symbol synchronization, di-demodulation or inverse suasisoherent using a first or second symbol sacrifice, deemodulasióp diresta. or inverse echo equalizer using a transverse equalizer (FIR), a decision feedback backfill equalizer, or a Viterbi equalizer, and differencial demodulation that insulates the displacement complex sansation due to time dispersion or those, 22, an ssnfor method The most important thing is the step of: demodulating signals received on a first or second antenna. using one of several algorithms. 23. A method of conformity is the ionic device 22 which further comprises the step of: demodulating an euma or weighted difference of signals resibidae in the first or second antenna. 24. An apparatus for reselection in Pi / 4-DQPSK signal diversity comprising a first and a second means d? resepsióp asoplados to a first antennas of resepsión to receive dishas signals, an analogue information converter means in digital asoplated to dishos first and second resepsion means to produce a first and a second streams of awkward numbers, and a signal processing means for processing said first and second streams of complex numbers to produce symbols demodulated comprising: a first means of a somplex multiplier to save a produsto of pairs of somplended numbers spaced by a DQPSK symbol among them in a first flow of somplite numbers, are the conjugation of one of disho pair to produce a flow of first produstos, a second medium of a sompletic mutant to salsulate a produsto of pairs of somplended numbers spaced by a DQPSK symbol among them in the second group of somplex numbers, in the conjugation of one of disho par to produce a flow of second produtos, a complex means of adding together the first and second corresponding products to produce In a flow of values that are combined in different levels, a complex means of correlating the flow of values in different levels is a pattern of sonoside symbols to determine a preferred ion, and a means of separation to evaluate values in a variety of sonority. they are disha sinsroni zasión preferred to extract bits of information. 25, An apparatus for diversity reselection of signals Pi / 4-DQPSK comprising: a first and a second means of resepsiop asoplated to a first and second receiving antennas to receive signals, an analog information converter means in digital asoplads a dishoe first and second receiving means for producing a first and a second stream of somplended numbers, and a signal processing means for processing said first and second sommelous streams to produce demodulated symbols that appear: a means of sorrels The somplejo ionator to correlate said first and second flows of somplended numbers is a pattern of sonoside symbols to produce a first and a second signal of sorrelation, a weighted admission means for adding first and second flow of weighted somplified numbers using first and second soefisientes of correlation to produce value sombipadoe in diversity, a means of What is the correlation between different values that are combined in diversity are the pattern of ssnosids symbols to produce a third soefisient? d? correlation, and processing means to process first, second and third correlation soefisient to select a demodulation algorithm. 26. An apparatus according to claim 15 further comprising a means for de-modulating minus one of the first flow of full numbers, second second flow of shaded numbers, and dish values combined in diversity using dishs selected demodulation algorithm. . 27, A method for transmitting information between a first and a second station comprising the steps of: sodifisar some part of disha inforraasión using a sodification algorithm d? correction of errors to obtain coded symbols, intercalar dishos sodicized symbols are symbols that represent a remainder of information not sodifisada between a first pattern and a second pattern of symboloseposidos to obtain groups of symbols for transmission, transmit dishos groups of eimboloe from disha first stasis hasia disha second stasis, reshaping disha transmitted signal, in disha second staion and produce a flow of numbers somplejos there, correlate this flow of sompl numbers are first and second patterns of symbols sonosidos to determine a first and second sinsronizasisnee corresponding to a maximum sorrelation and a first and a second values of sorrelation sorption, to proceed to the first and second qualities of sorrelation to determine a preferred method of desmsdulation between various methods of demodilization., demodulating said sodiplised and non-sodipized symbols using the preferred method of demodulation to obtain demodulated sodiplised symbols and demodulated non-sodi fi ed symbols, reprogramming scrambled codified symbols using an error corrector dessodi to obtain desmodified symbols, and combining said symbols. ssected ssdsdi symbols are disproportioned symbols not demodulated to ressinstitute disha sion, 28. A method to transmit infrasrmation between a first and a second staging that includes the steps of: scribing some part of disha sion using an error siphoning algorithm for get symbols sodifisados, intersalar dishos sodifisados symbols are simbicos that represent remains of non sodifised information and adding a pattern of sonosidos symbols to obtain groups of symbols for transmission, transmit dishoe group d? symbols from disha. The first station has a second station as a radio signal, the second radio signal is transmitted in the second station using a first antenna and a second antenna to produce a first received signal and a second received signal, to process the first and second signals received to obtain a first and a second stream of shaded numbers, correlating said first and second streams of discrete numbers is a pattern of sound symbols to determine a first and a second one if nsroni zas corresponding ions for a maximum sorrelation and a first and second corresponding sorrelation salinity values, first dish out and second sorrelation quality values to select one of first and second discrete number flows for desraodulasion using first and second synchronization dishas, respectively, and desmsdulating disho stream of selected sompl numbers to produce a number of ei The symbolization of demolition, symbolization of desmixed symbols corresponding to dish symbols, symbolizing and processing of scattered symbols using a dessodi fisasi? p method of sorroring errors in order to obtain systered data, and selecting desmsdulating symbols that correspond to disproportioned symbols of unsorted information for symbiosis are dishos dessodi symbols fisados to ressnstituir disha informasión.. 29. A method for transmitting information between a first and a second step comprising the steps of: sodipizing a part of the information using an algorithm for detecting errors in order to obtain symbols, interspersing dish symbols, symbols are symbols representing Non-sodifised information remaining between ssnosidos symbol patterns to obtain groups of symbols for transmission, transmitting groups of symbols from disha first station hasia disha second station in the form of radio signal, reshaping disha radio signal transmitted in disha second station using a first antenna and a second antenna to produce a first received signal and a second received signal, to process first and second received signals to obtain a first and second flows of superfluous numbers,, first to distinguish first and second flows of shaded numbers are dishwashing patterns of sound symbols for To determine a value of whether nsronizas ion for a maximum sorrelasión and a value of sorrelosión salor sorrespondiente for sada flow and pattern, to process dishosidad values of salidad to select one of dishos first or second flows of sompljos numbers, a sinsroni zas ion Preferred and a demodulation algorithm for producing demodulated symbols, select among said demodulated symbols those that correspond to dishwashed symbols and prosecute selected symbols using a decoding method. correction of errors to obtain decoded symbols that have a reduced probability of transmission error, and s? lessioned demodulated symbols that correspond to dishos informational symbols not sodifised for combination with dishos dessodi symbols fissured to reconstitute disha informas ion, 30, A system diversity radio receiver for desmodifying modulated radio information signals comprising: first and second antenna means connected to a first and second receivers to produce first and second amplified received signals; a means of converting analogue to digital information to sonar segments of first and second signals received amplified in first and second groups of numerical samples sorr? spondientee and almasener dishos group in a memory; means of proshaping asphalted to disha memory to salsulate from sample groups a characteristic indicative of signal salting for sada group of samples; a means of desision to use these signal quality characteristics to determine which of the sample groups should be flushed; and a means of dessodi fission to dissect the group of samples chosen by means of a half desicion to reproduce the information. 31. The system of radio resepsión in diversity of the resepsión 30, where: disho processing means also serves to combine first and second groups of samples to produce a group of samples and to show a characterist isa indicative of saliency of sada sign one of disho first group of samples, disho second group of samples and dishs group of sampled samples. 32. The diversity radio receiver system of claim 30 wherein: a processing means also serves to combine first and second sample groups using a number of predetermined weights to produce a corresponding number of groups of combined samples and to distinguish a characteristic indicative of signal output for one of the first group of samples, second group of samples, and group of sample samples. 33. A sonic reseptor system is the claim 30 where the mean of conversion of analog information in digital is a digi tal i zador d? diresta phase. 34. A reseptor system of sonification are the claim 30 where the means of conversion of apalógisa information into digital is a digi tal i zador d? sign sompleja logpolar. 35. A receiver system of ssnformidad are the rei vindisas i? N 31 where disho half of prosheamypto symbina samples of disho first and disho second groups of samples that were converted from analog to digital information at appropriate times. 36. A sonicity receiver system is the claim 31 wherein a medium of prosecution is used to sample samples from first and second groups of samples that were converted from analog to digital information in displaced times. 37, A sonicity receiver system is the ionic host 30 where disha indicative of saliency is a signal inteneity averaged over the signal segment, 38. A system of reception of sonicity is the rei vindisas ion 30 where disha signal signal output characteristic is one. correlation with one. signal form conosida insluido in disho signal segment. 39. A system of reception of sonformity are the claim 31 where disha characteristic of the saliency is a correlation are one. signal form sonoside insulated in disho signal segment. 40. A sonic resection system is the re vindisas ion 39 where dishoe groups of somated samples are formed by weighted symbiosis using dishas correlations to form complex weights. 41. A sonicity receptor system is the ion ion 31 in which samples of combined sample groups are formed by the mu ti ipl ies of samples of the first group taken from a dietary information symbol are suction of a sample and adicion of the produsts to a corresponding produst salted from the second group. 42. A system of resection of the sonformity is the vindication 41, where the disjoint product of the samples of the second group sorresponde a. a time offset for products from samples of first group, 43. A system of reception of sonicity is thedispersion 32 whereby the processing medium samples samples of first and second sample groups that were converted from analog information into digital in corresponding times, 44. A system of reception of conformity are the reisinistration 32 wherein disho processing means mixes corresponding samples of first and second group of samples that were converted from analog to digital information in displaced times. 45. A seven a of reception of sonformity are the vindication 42 where disha characteristic of saliency is a correlation are a form of signal sonosida insluida in disho eegmento d? signal. 46. A sonic resection system is the vindisas ion 45 in which dish groups of combined samples are formed by weighted combination using different correlations to form weighted weights. 47. An apparatus for the demodulation of signals modulated in the quaternary phase comprising: a means of sepsis for reshaping dishae modulated signals and providing a signal of intermediate frequency of hard limitation and an approximately lsgaritmisa signal intensity indigestion; a means for digitizing disha signal of intermediate frequency and disha indisasion of signal intensity in a stream of somplite numbers; an equalization means to demodulate the flow of shaded numbers to produce values of compensated real and imaginary symbols for multiple path distortion; and a means of correcting errors to process dishos values of real and imaginary symbols to reduce the number of transmission errors. SUMMARY OF THE INVENTION A method and system for demodulating signals received in radiocommunication systems is presented. The signals modulated by Pi / 4-DQPSK can be demodulated to provide additional quality measurements and to facilitate combination or selection in diversity. In testimony of which, I have signed the above description and novelty of the invention as attorney-in-fact of ERICSSON INC., In Mexico City, Federal District today, November 27, 1995. p.p. e ERICSSON INC.