MXPA98004883A - Circuit diversity combina - Google Patents

Abstract

This invention relates to a method and apparatus for the diversity combination of two electromagnetic signals (11, 21) within a receiver (25). Two separate antennas are used (10, 20). The antennas (10, 20) are separated by some combination of spatial separation, polarization and radiation pattern. The first antenna (10) receives the first signal (11) and the second antenna (20) receives the second signal (21). Coupled to at least one of the antennas (10, 20), there is a circuit (12) to vary the gain and the phase gives the signal (11, 21) received by said antenna (10 or 20 respectively). The gain and phase are forced to be selected from within a finite set of pre-selected discrete gains (6) searches through all gain and phase combinations to optimize the signal-to-signal deterioration ratio within the receiver . When the signal-to-signal deterioration ratio has been optimized, the gain and phase are fixed. The exploration (6) is divided into two discrete explorations, a relatively approximate exploration and a relatively accurate exploration. The scanning step (6) is typically performed when the signal-to-signal deterioration ratio drops below a preselected value. A scanning module (6) searches through all gain and phase combinations to optimize the signal-to-signal deterioration ratio within the receiver. When the signal-to-signal deterioration ratio has been optimized, the gain and phase are fixed. The exploration (6) is divided into two discrete explorations, a relatively approximate exploration and a relatively accurate exploration. The scanning step (6) is typically performed when the signal-to-signal deterioration ratio drops below a preselected value. An intelligent rounding module (14) can be used to recover the phase of the carrier of the combined signal (22) without any significant impairment of performance.

Description

"DIVERSITY OF COMBINED CIRCUITS" DESCRIPTION OF THE INVENTION Technical field This invention is related to the technique of combining electromagnetic signals, coming from separate antennas, in order to cancel the common channel interference and mitigate the selective frequency fading. The following references are cited in this specification, using the following reference numbers: [1] D. C. Cox, "Universal Digital Portable Radio Communications" (Universal digital communications using portable radios). IEEE Proceedings, Vol. 75, No. 4, pages 436-477, April 1987. [2] R. C. Bernhardt, "User Access Portable Radio Systems in a Co-channel Interference Environment "(User access in portable radio systems in a common channel interference environment), IEEE Journal on Selected Areas of Communications, Vol. 7, No. 1, pages 49-58, January 1989. [3] PB Wong and D. O Cox, "Low Complexity Co-channel Interference Cancellation and Macroscopic Diversity for High Capacity PCS" (Canceling low complexity common channel interference and macroscopic diversity for personal communications systems (PCS) of high capacity), Conf. Re8rd IEEE ICC '95, Seattle, WA, pages 852-857, January 18-22, 1995. [4] TR-INS-001313. "Generic Criteria for Version 0.1 Wireless Access Communications Systems (WACS) "(Generic criterion for version 0.1 of communication systems with wireless access), Bellcore Edition 1, October 1993, revision 1, June 1994. [5] D. C. Cox, "Wireless Personal Communications: What Is It" (Wireless Personal Communications: What are they?), IEEE Personal Communications Magazine, pages 20-35, April 1995. [6] D. E. Thomas and P. R. Moorby, 'The Verilog hardware description language' (Verilog Hardware Description Language), Second Edition, Boston: Kluwer Academic Pub., 1995. [7] E. Stemheim et al., "Digital Design and Synthesis with Verilog HDL" (Digilat Design and Synthesis with Verilog HDL), San Jose, CA: Automata Pub. Co., 1993. [8] J. C.-l. Chuang and N. R. Sollenberger, "Burst Coherent Demodulation with Combined Symbol Timing Frequency Offset Estimation and Diversity Selection" (Coherent burst demodulation with an estimate of symbol synchronization frequency offset combined with diversity selection), IEEE Transactionas on Communications, volume 39, number 7, pages 1157-1164, July 1991. [9] J. C.-l. Chuang and NR Sollenberger, "Burst Coherent Detection with Robust Frequency and Timing Estimation for Portable Radio Communications" (Consistent burst detection with a strong frequency and synchronization estimate for communications with portable radios), IEEE GLOBECOM Record Conf '88, Hollywood, FL, pages 804-809, nov. 28 a.c. 1 of 1988.

[10] J. C.-l. Chuang, "The Effects of Time Delay Spread on Portable Radio Communications Channels with Digital Modulation ", The effects of the delay distribution in portable radio communication channels with digital modulation), IEEE Journal on Selected Areas in Communications, Vol. Sac-5, No. 5, pages 879-889, June of 1987.

The present invention consists of a method and apparatus for the combination in diversity of two electromagnetic signals (11, 21) within a receiver (25). Coupled to the receiver (25) there are two antennas, a first antenna (10) to receive the first signal (11) and a second antenna (20) to receive the second signal (21). Coupled to at least one of the antennas (10, 20) there is a circuit (12) for varying the gain and phase of the signal (11 or 21) received by said antenna (10 or 20). Gain and phase are constrained to be selected from within a finite set of pre-selected discrete gains and a finite set of preselected discrete phases.

BRIEF DESCRIPTION OF THE DRAWINGS These and other more detailed and specific objects and features of the present invention are described in the following specification, with references to the accompanying drawings, in which: Figure 1 is a block diagram illustrating the combination in microscopic diversity of the present invention using two receiver antennas 10, 20.

An example of a set of complex weights A used for the diversity combination is illustrated in Figure 2. The complex weights A are used in the approximate scanning step and are represented by the hexagonal markers, while those used for a particular example of the precise scanning step are represented by shaded rectangular markers.

Figure 3 is a set of three graphs, related in time, showing the signals s, S and S '. The signal S is delayed by 1 sampling period to change its phase at an angle of -45 ° to S '. The signals s and S 'are then in phase and can be suitably added.

Figure 4 is a process flow diagram illustrating the diversity combination of the present invention in a receiver 25. The functional units enclosed in the solid boxes are unique in the present invention.

Figure 5 is a functional block diagram showing the data path circuits for the primary, secondary and reserve (relatively low speed) states of phase recovery of the present invention (the components enclosed in the rectangular line box of points 29).

Figure 6 is a block functional diagram showing the parallel circuits of data path (relatively high speed) for the exploration of the complex optimum weights A for the combination of signals and for the optimal synchronization of corresponding symbol.

Figure 7 is a synchronization program for Ta phase recovery 29 and for exploratory states 6, with their corresponding clock frequencies in a PACS TDMA frame period. It is assumed that the data is received in the first time interval, T1 (the shaded area).

Figure 8 is a set of two graphs showing a simulated quality of normalized signal (y) vs. SIR (acronym for "Signal Interference Ratio") in a limited interference environment.

Figure 9 is a set of three graphs showing the results of the computer simulation of the hardware for the two receiver antennas 10, 20 located in the statistically worst area. Macroscopic diversity is incorporated based on the channel power measurements. Cs, n and s are 8, 4 and 10 dB, respectively. Note that the points Y, '* and' o 'are the calculated simulated results of the signal carriers only. The continuous curves are the results obtained by measuring the SQ (acronym of "Signal Quality" or signal quality) in the hardware simulations.

Figure 10 is a set of two graphs showing the simulated and estimated average value of the SQ vs. signal quality. the SIR relationship in a limited interference environment. The values in the SQ axes depend on a specific hardware implementation.

Figure 11 is a set of three graphs showing the results of the computer simulation of the hardware for the two receiver antennas 10, 20 located in the statistically worst area. The macroscopic diversity is incorporated based on the channel power measurements. Cs, n and s are 8, 4 and 10 dB, respectively. Note that the points Y, '* and' o 'are the calculated simulated results of the signal carriers only. The continuous curves are the results obtained through the use of SQ by the hardware.

Figures 12A and 12B are the two power delay profiles used to study the effectiveness of frequency selective fading mitigation using diversity selection and diversity combination. The initials DS come from the English "Delay Spread" or distribution of the delay. T is the time interval between the channel pulses. Pi is the variation of a complex Gaussian random process for the impulse response of the channel in the period of time T.

Detailed description of preferred variations 1. Introduction Figure 1 illustrates the basic operation of a receiver 25 using the present invention. There are two separate antennas 10, 20 coupled to the receiver 25. The antennas 10, 20 are separated by some combination of spatial separation, polarization and radiation pattern. A first signal 11 is received by the antenna 10. The signal 11 is a combination of the desired signal S and one or more interference signals l-¡, l2, etc. Similarly, the signal 21 received by the antenna 20 is a combination of a desired signal s and one or more interference signals, i2, etc. A gain and / or attenuation circuit 12 is inserted into at least one of the two branches 61, 62. The circuit 12 introduces a complex weight A into the circuit. A has a gain component (k) and a phase component (Q). The two branches 61, 62 are combined in the combiner 15, forming a combined signal 22.

In a PCS high-capacity personal communications system (PCS stands for "Personal Communications System"), for a given bandwidth, the common channel interference CCl (CCl stands for "Co-channel") Interference ") limits the capacity of the system [references 1, 2]. Usually, the CCl value is dominated by a common channel interference due to the phenomenon of the screen effect having a local average value, normally logarithmically distributed, of the power of the received signal. The diversity combination of two antennas 10, 20 (Figure 1), as described herein, is capable of canceling this dominant interference and producing an improvement in the signal to SIR interference ratio of at least 3.8 dB over the selection diversity. conventional two-antenna [reference 3]. An improvement of 3.8 dB in the SIR corresponds to an increase in the capacity of a TDMA / FDMA wireless system by a factor of 1.5. This improvement of the system is obtained with the constraint that the complex weight (A in Figure 1) must be a chosen value from the following Equation 1, with k = 3 and? Q = 18 ° [reference 3]: 3A3 = 2n , n = -k,. . . , k - 1 or 2n + 2n + 1, n = -k,. . . , k - 2 QA = m *? Q, m e. { Z 30 ° £ m *? Q < 360 °} (equation 1) where k is a positive integer and Z is the set of all other integers. Other values of 3A3 and QA can be used, with a smaller number of options resulting in lower performance and a greater number of options resulting in higher performance.

The present invention uses low complexity methods and circuit architectures for diversity combining without requiring training sequences or reference signals on the radio communications links. The PACS standard [references 4, 5] of low-level PCS in the United States is used herein to illustrate the present invention. However, the performance of any frequency reuse system, for example, a PCS system, can be improved by similar methodologies of combination. The methods and circuit architectures described here have been verified in the Verilog hardware description language [references 6, 7]. The results of hardware simulation show that there is a reduction of less than half a dB of the optimal solution, taking into account the constraint that A must satisfy Equation 1 with k = 2 and? Q = 45 °. As indicated above, smaller increments of 3A3 and of Q reduce signal deterioration, while larger increases result in further deterioration.

Currently, PACS uses a selection in diversity without selective matching and has an irreducible average of WER (abbreviations in English of "Word Error Ratio" or relation of errors of words) of approximately 2.4% when the distribution normalized of the delay in the channels of radius is approximately 0.16 [reference 8]. The combination of diversity without self-adapting equalization, but using the methods and circuit architectures described in this patent application, can result in the same irreducible average of WER when the normalized distribution of the delay is 0.3, which corresponds to a delay distribution 1.6 ms for a symbol period of 5.2 ms. This can extend the range of the radius over which PACS can operate.

These improvements in the system can be obtained with only a slight increase in the signal processing complexity of two linear receivers 25 with automatic gain controls, 6-bit analog-to-digital conversion 2 and a larger number of digital circuits to implement the methods of combination of signals. If smaller increments are used in 3A3 and? Q to improve performance, as indicated above, then more bits may also be needed in the analog-to-digital converter 2.

We describe the optimization method in Section 2 below under the constraint stipulated in Equation 1. Section 3 introduces some aspects associated with the implementation and with our circuit architectures for the diversity combination of the two antennas 10, 20. The Section 4 presents the results of the hardware simulation for the cancellation of CCl and the mitigation of frequency selective fading. 2 Exploration method In this section we describe a simple but effective method to find the optimum A complex weight for the combination in diversity. This method requires fewer calculations than for the thorough exploration described in reference [3J. No training sequences or reference signals are required on the radio communication links to search for optimal A complex weights.

With the gain and phase increment quantified in Equation 1 as k = 2 and? Q = 45 °, respectively, the combination results in diversity are within a few tenths of a dB of the optimal combination using continuous gain values and phase [reference 3]. Figure 2 shows the complete set of complex weights A quantized horizontally with phase and vertically with magnitude. The magnitudes and phases for this game representative of weights A are chosen for an easy implementation in circuit 12. One way to determine the weight compiejo To optimum for combination in diversity is by means of a meticulous exploration. A more conservative computing process is to evaluate weights A in approximate steps as represented by the shaded hexagonal points in Figure 2. For each amplitude and phase step, the SQ quality of the signal is calculated 24 for the combined signal 22. The quality of the signal can be defined as the average aperture of the ocular pattern in a sudden signal increase and is a good measure of the signal-to-signal deterioration ratio in the optimal symbol synchronization. Reference [8] describes this measurement of SQ. Then we can analyze the complex weights A that surround the hexagonal complex weight A that provides the best SQ, in the search for a complex weight that produces a better SQ.

Even though the values of SQ and SIR are used in the specification as examples of signal quality measures, it should be understood that any type of signal-to-signal deterioration indication may also be used. "Deterioration" includes, but is not limited to, factors such as noise, common channel interference and intersymbol interference.

In the example indicated by the shaded area of Figure 2, if the complex weight A with a magnitude and phase equal to 1 and 90 °, respectively (marked b2), provides the maximum quality of the SQ signal for the approximate exploration on the points Shaded hexagons, then their surrounding weights A, represented by the shaded rectangular points, are scanned to find the best SQ. The desired complex A weight, used for the combination in diversity, is the one that produces the best SQ over the complex A points that have been explored. This methodology of splitting the exploration into a relatively rough exploration and relatively accurate exploration reduces the possibility of ending up at a local maximum instead of a global maximum and eliminates the need to have to explore all the A weights. In Section 4.2 we show that this exploratory methodology results in very little depletion of the signal when compared to a thorough exploration. However, this exploration method covers only 20 complex A weights, instead of all 56 A complex weights. 3 Implementation and circuit architecture The circuit architecture described in this example variation is designed to be used in a PACS receiver 25 with handset [reference 4J. A base station receiver 25 can use the same techniques to achieve a comparable improvement in the diversity combination. More parallelism and circuit complexity will be needed in a base station receiver 25 to process all 8 bursts of data that need to be received. Receivers 25 may be associated with transmitters to form transceivers. A group of similar transceivers are located geographically so that they can communicate with each other using radio frequencies. When the transmitting portion and the receiving portion 25 of the transceiver operate on the same frequency, similar circuits and methods can be employed for the transmitter and the receiver 25. For example, the transmitter can transmit on the two antennas 10, 20 using a conjugate of the phase determined by receiver 25 in scanning step 6. This would be an effective strategy for transceivers with noise limiters, but would not be optimal in all cases for limited situations of common channel interference.

Most of this specification describes the circuits and methods used in conjunction with the receiver 25. These techniques can be adapted by the receivers to be used with other radio link architectures and modulations, as will be obvious for a person with sufficient knowledge of the job. 3. 1 Asymmetric sampling rates in two branches 61, 62 One way to adjust the phase difference between the two received signals 11, 21, for the diversity combination in a digital circuit 25, is to oversample the the signals 11, 21 and then delay a set of samples for one of the branches (61 or 62) in relation to the samples of the other branch (62 or 61). To be able to introduce a minimum phase difference of 45 ° (without interpolation between the samples) between the samples of the two receiver branches 61, 62, the signals received from a branch (61 or 62) need to be sampled at a speed of 8 times the IF intermediate frequency (IF stands for "Intermediate Frequency" or intermediate frequency). The signals received from the other branch (62 or 61) can be sampled at 4 times the IF, as in the old implementation of PACS [reference 8]. Using a sampling rate of 4 times the IF provides a convenient form of coherent demodulation of the data burst without the need for any training sequence [reference 8]. The phase change between the branches 61, 62, using elements of adjustable delays 7 in both branches 61, 62, is between -90 ° and + 90 °. The other half of the required phase changes (from + 90 ° to + 270 °) can be obtained by negating the samples of the signals received 11, 21 from one of the branches (61 or 62). Figure 3 shows a phase change of -45 ° by delaying the samples of the signal s of one branch (61 or 62) by 1 sample, with respect to the samples of the signal S of the other branch (62 or 61). ).

Other phase adjustment increments,? Q, can be produced by sampling the branches 61, 62 at different oversampling rates. A larger oversampling may result in smaller increments? Q, while a lower oversampling will result in larger increments.

In Figure 5 a variation of the present invention is shown in which each branch 61, 62 has its own gain attenuation circuit 12, which consists of a time delay element 7 (such as a series of engaging elements) and a gain / attenuation element 39, such as a stepped register.

The IF used in a PACS implementation is 4 times the SF symbol frequency (SF stands for "Symbol Frequency" or symbol frequency). Therefore, the maximum relative delay introduced by the sample delays for the combination in diversity is only one sixteenth part of a symbol period. That means that the maximum normalized delay distribution, introduced by the diversity combining process of the present invention, is only 0.031 for the flat fading radio propagation channel. Below, it will be shown, in Section 4.2, that this interference between additional symbols causes negligible deterioration in the overall performance of the system. For channels with selective frequency fading, general intersymbol interference is currently reduced by the diversity combination of the present invention (Section 4.3), even though the combination process introduces a small delay distribution in the flat fading channels. 3. 2 System overview An overview of the process flow in a receiver 25 is shown in Figure 4. The functions of each individual block are described in this subsection. The circuit architecture designs of the functional units, enclosed in solid striped boxes, are presented in the following subsections. Reference [4] presents the design and performance of the other functional units enclosed in boxes with dotted lines, except the signal comparison unit 3, which is a conventional circuit that only requires subtracting two numbers.

The signals received from the two antennas 10, 20 sdh processed first by the radiofrequency input circuit 1 to amplify and convert the signal to an intermediate frequency and to withdraw the adjacent channel signals. The signals are then oversampled and quantized by an analog-to-digital converter 2. The quantized signals of the two receiver branches 61, 62 are combined by the combiner 15 and the carrier phase is recovered in the primary phase recovery circuit PPR 31 (PPR stands for "Primary Phase Recovery" or primary phase recovery). The PPR circuit 31 uses the complex weight A and the symbol synchronization found / used in the previous received data burst. The Complex weights and symbol synchronization previously used are not initially available. Initially complex weights A and random character symbol synchronization are used in the PPR unit 31. After effecting the PPR using complex weights A and random character symbol synchronization, circuit 31 will probably trigger scan status 6 to find the weight complex A and the optimum character symbol synchronization for the initial group of data.

The SQ signal quality of the recovered phase is then compared to a predefined signal quality threshold, called (SQ *), in the comparator 3. SQ * is selected to provide a word error ratio of an acceptable value. If SQ is greater than SQ * the recovered phase is coherently demodulated by the demodulator 4. Otherwise, the SS scan state is activated (SS stands for "Search State"). the coherent demodulation, the recovered bits are examined 5 to determine if they have errors and to see if the color code of the common channel is correct [reference 3]. If both checks are correct, the burst of recovered data is transferred to output 9. Otherwise, the scan status SS 6 will be triggered.

The scanning state 6 uses the method described in Section 2 to scan and determine the optimal complex weights A for the combination of signals. The symbol synchronization corresponding to both optimal complex weights A and the negation of those weights A are also found in SS 6. "Symbol synchronization may be different for different complex weights A used in the diversity combination process. secondary phase recovery SPR 32 (SPR stands for "Secondary Phase Recovery") uses the optimal A complex weight and symbol synchronization provided by SS 6 to combine the received signals 11, 21 and recover the phase of their carriers. The recovered phase is then coherently demodulated at 43. If the recovered data burst contains some error, as detected by the error inspection module 51, a word error 8 is declared and the recovered burst is bounced. If I dont know detects no error, module 52 will check the color code. If the color code is similar to the desired one, the recovered burst of data will be transferred to the output 9. Otherwise, the phase recovery backup circuit BPR 33 (BPR stands for "Backup Phase Recovery" or backup phase recovery) will be triggered. A burst of data that is error-free, but color-code verification has failed, will likely be a dominant interference (I or i in Figure 1), indicating that the burst has been enhanced, but that the desired signal [S os] has been reduced by the diversity combination, using the negation of the optimal complex weights A for the combination of signals can partially cancel this dominant interference and add the desired signals [S, s] in a constructive way. BPR circuit 33 uses the negation of the optimal complex weight A and the corresponding optimum symbol synchronization to recover the phases.The recovered phases are then coherently demodulated by the demodulator 44 and examined to detect errors and verify the color code by the module 53. If both exams are successful, the data burst is transferred to output 9. Otherwise, the received data burst is thrown at 8. 3. 3 Main blocks of the system All the bit lines drawn in the figures and the architectures discussed in Section 3 have been verified in the Verilog hardware description language. 3. 3.1 Combination of signals and automatic digital gain control In order to preserve the amplitude information for the diversity combination, it is not possible to use the hard limiter and the 4-bit signal quantization used in the original PACS 25 receiver. We have carried out Monte Carlo simulations to study the effect of quantization noise in the operation of the system. Using a 5-bit signal quantization, there is a deterioration of approximately 0.3 dB in system performance for diversity combining when amplifiers are used linear with automatic gain controls (AGC stands for "Automatic Gain Control" or automatic gain control) perfect. To relax the requirement of a perfect AGC, an additional bit can be used for the quantization of the signal. If 6 bits are used for the quantization of the signal and the set of complex weights A shown in Figure 2, the combination of the received signals 11, 21 of the two receiver branches 61, 62 can be carried out with a circuit of addition of 7 bits 3-2 15, a complete circuit of addition of 7 bits 15 and some elements of delay (hooks) 7. As indicated above, combining with a greater or lesser deterioration of the signal will result in more or fewer bits of signal quantization, respectively.

After the combination of signals, the amplitude of the combined signal 22 may be large when the two signals 11, 21 are mainly used to increase the desired signal strength, or small when the two received signals 11, 21 are used mainly for cancel the dominant interference. The condition that occurs will depend on the phases and relative amplitudes of the signal S, s and the interference I, i on the two branches. Therefore, we need a "digital AGC" after the combination of signals, so that both cases can be represented by the same number of bits, while avoiding the waste of hardware resources by carrying precedent bits of "zero" value.

The relationship between the quadrature component and the in-phase component determines the phase of the received signal 22. If the preceding bits of both components are "zeros" after the combination of the signals, we can move the bits of both values to the left a total of two spaces or registers, until the most significant bit MSB (MSB stands for "Most Significant Bit" or one of the two values) is / are 1. Then the two components can be divided into two 5-bit values. This is known as intelligent rounding (module 14 in Figure 5). It is essentially equivalent to a "digital AGC". Using intelligent rounding we can round a value of 9 bits to a value of 5 bits for each component, without lose any meaning in the recovered phase of the carrier, and we can dramatically reduce the amount of memory for the phase determination frame (by a factor of 2 *).

Even when QPSK modulation has been used to illustrate the operation of this invention, it should be understood that any type of modulation, including analog modulation, can be used. When analog modulation is used, A / D converters 2 are still used. 3. 3.2 Decoder circuits Figure 5 shows the main data path components 29 for the primary, secondary and backup data recovery states (PPR 31, SPR 32 and BPR 33, respectively). The signals received from a branch 61 are sampled at 6,144 MHz (8 * IF) by the switch 35 (1) and then are quantized at 6 bits, for each sample, by the A / D converter 2 (1). The signals received from the other branch 62 are sampled at a lower frequency of 3072 MHz (4 * IF) by the switch 35 (2) and quantized to the same bit precision by the A / D converter 2 (2). Alternatively, the same A / D converter 2 could be used for both branches 61, 62 in combination with sampling and clamping circuits, as is well known in the art.

The quantized samples received are stored in three memory blocks of 744 bytes each, named XO-RAM 40, X1-RAM 41 and Y-RAM 42. 744 bytes = 6 bits * 16 samples per symbol * (60 + 2) symbols / 8 bits. Additional storage of 2 symbols is used as a backup, because there may be a change in symbol synchronization between received bursts. Samples taken at the highest sampling rate are sandwiched between memories XO-RAM 40 and X1-RAM 41. Samples stored in these three memory blocks 40, 41 and 42 are used in scanning states 6, SPR 32 and BPR 33. State PPR 31 uses samples taken directly from the outputs of the two analog to digital converters 2. This is desirable because the use of memories 40, 41 and 42 is avoided, which saves energy.

As the optimal symbol synchronization is supplied to the phase recovery states, this data path circuit 29 calculates only a phase value of 8 bits per symbol. This is different from the design in reference [8], where the optimal symbol synchronization is determined for each received burst and, therefore, the phase recovery circuit needs to calculate 16 8-bit phase values for each symbol.

The processes implemented in the component generator I and Q 13, in the phase recovery circuit 30 and in the frequency deviation estimator 16 are described in reference [8]. The frequency shift estimator 16 measures the signal quality, provides coherence and combats the frequency inequality in the oscillators of the receiver 25 and the transmitter for this QPSK variation. The processes implemented in the signal combiner 15 and in the intelligent rounding module 14 were presented in the previous subsection. The functional units, shown outside the rectangular box with dashed lines 29, are not discussed in detail in this specification and are encoded in the computer language C using floating-point calculations for the general simulations of the system described below. 3. 3.3 Exploration circuits Figure 6 shows the components 6 of the data path circuit used in the exploration of the complex weight A optimal for combination in diversity and to find the optimal synchronizations of the symbol, corresponding both to the optimal complex weight A and its negation. The scanning circuits 6 described in Figure 6 contain three main blocks of parallel circuits. If time is not a problem, a block or two blocks can be used in parallel. If time becomes a problem, a larger number of circuits can be used in parallel. There is an interaction between the complexity of the circuit and time. Alternatively, one of the blocks can be used to satisfy the same time constraint if it has a faster synchronization period, as will be obvious to anyone with knowledge of the trade.

The memory blocks 40, 41 and 42 are the same as those shown in Figure 5. The signal recovery and phase recovery combiner blocks 23 are basically the same as those used in the decoding circuits described above, except that they are more in line and the recovered phase is a value of 6 bits instead of 8 bits.

This 2-bit reduction in the representation of the phase value is due to the fact that only the least significant 6 bits of a recovered phase are needed for the SQ signal quality and for the symbol synchronization measurements by the modules 24. The methods and circuit architectures for SQ measurements and symbol synchronization 24 are discussed in more detail in reference [8]. As indicated above, SQ is just one example of the signal-to-signal deterioration relationship. Therefore, in the general case, the circuit 24 is a signal-to-signal deterioration and symbol synchronization measuring circuit.

The scan circuit 6 uses the methods discussed in Section 2 to explore and obtain the optimal A complex weights. Throughout the approximate scan (represented by the hexagonal points shaded in Figure 2), one has access to Y-RAM 42 and only one of the memories X-RAMs 40, 41. During each access to memory, this circuit 24 Measures three different SQ values and ST symbol synchronization (ST stands for "Symbol Timing") for three different A complex weights. These three values SQs and STs pass to the signal quality comparator 3, which compares the three SQs, selects the highest SQ and passes it to the highest SQ register 17 and the addition circuit 18. The comparator 3 also passes the ST associated with the highest SQ to the ST register 19.

After a total of four accesses to the memory for the approximate scan step, one has access once more to the X-RAM memories 40, 41 used in the approximate scan step and the Y-RAM memory 42 for the precise step of the scan (represented by the shaded rectangular dots shown in Figure 2 for a particular case). Next, the other memory X-RAM 41, 40 and the memory Y-RAM 42 are accessed twice for the precise passage of the scan. That is, the exploration circuits need to operate only 7 times to find more than 20 different complex A weights in order to find the desired game for the combination of signals. All of the signal quality comparisons of these scanning circuits 6 are sent to a control circuit unit (not shown) where the desired complex weight A is determined (the weight that produces the highest SQ signal quality). Note that the complex weights A used in the approximate step of the scan are always the same for any received burst of data, but the complex weights used in the precise step of the scan may be different for each individual burst.

The optimum symbol synchronization corresponding to the optimum complex weight is the final value stored in the symbol synchronization register (ST). When the optimum symbol synchronization is determined for the negation of the optimal complex weight, which is needed in the BPR 33 state, only one third of the scanning circuits will be put into action. The content of the register 19 passes to the RAM memories 40, 41 and 42. This information tells the RAM 40, 41 and 42 when the signals must pass to the combiner 15. 3. 4 Real-time programming for burst demodulation A TDMA frame period for a PACS radio communications downlink is shown in Figure 7. The following terms are used in Figure 7: KEY: PPR - primary phase recovery 31 IF - intermediate frequency (768 kHz) SPR - secondary phase recovery 32 SF - symbol frequency (192 kHz) BPR - backup phase recovery 33 ST - symbol synchronization SS - scanning state 6 DCW - desired complex weight A X - signal quality comparison 3, coherent demodulation 4, error checking and color coding 5 Y1 - coherent demodulation 43, error checking and color coding 51, 52 Y2 - coherent demodulation 44, error checking and color coding 53 A frame period of 2.5 ms, for example, a frame rate of 400 Hz, is divided into 8 individual bursts. The processing time has been selected to ensure that the signals received after the diversity combination are decoded within a frame period after receiving the first sample of the desired burst, while maintaining the maximum clock frequency used in the frame. circuit at the maximum speed of 6.144 MHz (8 * IF). The speech transmission delay is increased by approximately 2 ms, compared to the original diversity selection PACS receiver. (This delay could be reduced by using higher clock frequencies, which could increase energy consumption a bit).

A frame period is subdivided into several different states of phase recovery. As shown in Figure 7, each state uses a different clock frequency, except the secondary phase 32 and backup 33 recovery states, which use the same frequency. For a fade rate of 3 Hz, which is 3/4% of the frame rate, or less, the circuit drives the primary phase recovery state (PPR) 31 only most of the time. Because only one phase value per symbol is calculated in state PPR 31, most of its data path circuit components operate at a clock frequency equivalent to the symbol frequency (192 kHz). The control circuit units always operate at 8 times the intermediate frequency (8 * IF = 6.144 MHz), that is, at the maximum sampling frequency of one of the two branches.

There is a burst period, called X, between the state PPR 31 and the state of scan 6, and also after the other phase recovery states (called Y1 and Y2). These periods of time are reserved for coherent demodulation 4, 43, 44, verification of errors 5, 51, 53 and verification of color code 5, 52, 53. These functional blocks are conventional and, therefore, do not they are discussed in detail in this specification. They will be detailed in reference [4].

In order to conserve energy, the scanning state 6, which operates at 8 * IF = 6.144 MHz, is only activated if the signal quality (SQ) measured in the PPR 31 state falls below a predefined threshold of SQ ( SQ *), if there is an error in the decoding or if the color code verification fails. The scanning state 6 will determine the optimum A complex weight using the scanning method described in Section 2 and its corresponding symbol synchronization for the diversity combination in the secondary phase recovery state (SPR) 32, which operates at the frequency intermediate (IF).

The handset receiver 25 drives the backup phase recovery, state 33, using a clock frequency equal to the IF, only when the color code fails 52 in the SPR state 32. The BPR state 33 uses the complex weight negation An optimum found in the scanning state 6 and its corresponding symbol synchronization determined immediately after the BPR 32 state (as shown in Figure 7) for the combination of signals and phase recovery. The symbol synchronization for state BPR 33 uses the third part of the SS 6 circuits, as indicated in the previous subsection.

Hardware simulations and results This section describes the radio channel simulation and modeling procedures for our computer simulations and presents some of the results of system evaluation. The simulations were carried out both for the case dominated by the common channel interference and for the case dominated by the intersymbol interference ISI (ISI stands for "Intersymboi Interference" or interference). intersymbol). 4. 1 Simulation procedures The radio channel models, both for the limited cases of common channel interference and ISI, are discussed in their own subsections. This subsection will describe the simulation procedures that are common for both cases.

The stage 29 recovery phase and scanning stage 6 for the optimum A-complex weight were coded in the Verilog hardware description language with the corresponding programs of synchronization, clock frequencies and finite number of bits for computing as discussed above. The rest of the functional units necessary for the simulation of the system was coded in computer C language, using floating point calculations.

The AWGN Additive White Gaussian Noise (AWGN stands for "Additive White Gaussiam Noise "or additive Gaussian white noise) was not included in the simulations.The only noise in our simulations was the quantization noise due to the finite number of bits used in signal quantization and hardware integer computation (shown in Figures 5 and 6).

Quasi-static channels were assumed in the simulations. That is, it is assumed that the channel is constant through a single burst. Below is a list of the other conditions that were used in the computer simulations: Data transmitted independently, randomly, for each base station and each individual burst.

QPSK modulation p / 4 with differential coding and Gray coding [reference 4J.

Spectral form of elevated cosine of Nyquist with a progressive attenuation factor of 0.5 [reference 4].

• Symbol synchronization distributed evenly in a symbol period (not restricted to the 16 sampling points in a symbol).

Uniformly distributed frequency deviation between ± 11% of the symbol speed (corresponding to 10.6 ppm for a 2 GHz radio frequency).

Linear amplifiers with AGC in the receiver.

Quantification of 6-bit signal.

Second order phase interlock circuit with phase loop gain of 2-2 and frequency loop gain of 2 ^ [reference 9]. 4. 2 Cancellation of common channel interference The following conditions, in addition to those described in the previous subsection, were used in computer simulations for the cancellation of common channel interference. 8 channel sets (Cs = 8). • Square cell geometry. Close only of the common channel base stations 25 only. Asynchronous and random transmission times between the base stations 25. • The handset 25 located in the worst area from one point of view statistical. The statistics of the received signal are distributed in Rayleigh form and normal logarithm. Exponent of trajectory loss of 4 (n = 4) in propagation d. **. Standard deviation of the normal logarithmic deviation of 10 dB (s = 10 dB).

Reference [3] contains detailed descriptions of the simulation procedures and these conditions.

Before being able to compare the results of the hardware simulations with the results of the computer simulations in the reference [3], it is necessary to relate the measurement of the average of SQ with SIR in the limited environments of interference. To do this, we first generated 8 carriers of random signals of common channel with independent statistics of normal logarithm and Rayleigh. The desired signal carrier was then generated and scaled appropriately to produce the required SIR. Then random streams of data were added to these carriers. Each received burst had a different set of carriers and data streams. The SQ signal quality of a burst was then measured at the output of the phase recovery circuit without diversity and using the Verilog simulations. The normalized average of the SQs [reference 8] was more than 20 bursts and a total of 40 bursts were plotted, according to the SIR, in Figure 8.

In Figure 8 it is shown that the normalized mean SQ converges reasonably well using only 20 bursts for the SQ measurement in the region where SIR is greater than 6 dB. This includes the 8 dB to 20 dB region, in which we are most interested. Below 8 dB, the error ratio is too large to be used, whereas, above 20 dB, the error ratio becomes essentially zero. This figure is very similar to Figure 5 presented in reference [8]. Particularly, when SIR and SNR are greater than 15 dB, these two figures are almost the same. The The difference between the two figures is mainly attributed to the different signal impairments introduced in computer simulations: AWGN is used in reference [8], while common channel interference is used for the results presented in Figure 8.

For the next step in the simulation process, 1483 independent bursts were generated with macroscopic diversity selection based on power measurements. This is the same case as that presented in Figure 4 in reference [3], except that here k = 2 and? Q = 45 ° are used in equation 1, instead of k = 3 and? Q = 18 ° in reference [3]. The SIRs are calculated for these 1483 bursts, both of the signal carriers only as described in the reference [3], and of the SQ measurements in the hardware simulations, for a single antenna, for selection in diversity and for combination in diversity The measured SQ is delineated in the SIR using the path shown in Figure 8 for the case of 40 bursts and using linear interpolation between the data points. The simulated cumulative distributions of SIRs are shown in Figure 9. The continuous curves are from the SQ measurements in the hardware simulations, and the discrete data points are from the calculation of the SC signal carrier (SC stands for "Signal Carrier" or signal carrier).

The results of the SQ measurement are in agreement with the results of the SC calculation for SIRs between 8 dB and 20 dB, which constitutes the region of interest for the cases of a diversity selection antenna. The slight difference is due to the uncertainties of the layout (the SQ to SIR plot is a process involving average values) and the imprecisions of the layout itself (the use of linear interpolation between the data points). For the same SIR range, the diversity combination of the present invention experiences a small deterioration (less than half a dB), compared to the results of the SC calculation, due to the following reasons: The ISI is introduced into the hardware simulations by moving the received samples to obtain the phase required for the combination of signals.

There is no frequency shift in the SC calculation and, therefore,? Q = 45 ° is an exact value. Due to the frequency shift in the hardware simulations,? Q is not exactly 45 °.

In the SC calculation a meticulous exploration is used, guaranteed to find the optimal solution. In the hardware simulations, the exploration method described in Section 2 is used.

There is a growing discrepancy between the hardware simulation results based on the SQ measurement and the SC calculation results for SIR less than 6 dB and greater than 20 dB for all three cases. This is because the SQ is saturated at both high and low SIR values and also because other different effects, such as quantization noise, become significant at low SIR.

Because the SQ signal quality is also estimated by the hardware during the frequency offset estimate, we change the SQ measured by floating point calculation of the phase recovery circuit output by the SQ estimated by the hardware using precision bit much smaller to repeat these same studies. Figure 10 shows a plot of the mean SQ value estimated by the hardware vs. SIR. The values on the SQ axis depend on the specific hardware implementation. Cumulative simulated distributions of SIRs, using the value estimated by the SQ hardware for plotting purposes, are shown in Figure 11. This Figure also shows that combining in diversity in hardware simulations causes small deterioration (less than half a dB), compared to optimum solutions, under the imposed constraint of the quantified complex weight.

For SIR values between 8 dB and 20 dB, both Figure 9 and Figure 11 are in accordance with Figure 4 presented in reference [3] for the cases of a single antenna and diversity selection. For this region of SIR, the combination in diversity of the present invention has lower deterioration (0.5 + 0.3) dB when compared to the results indicated in reference [3]. The further deterioration of 0.3 dB is due to the use of k = 2 and? Q = 45 ° in Equation 1 of the hardware simulations, instead of k = 3 and? Q = 18 ° that were used in the reference [3 ] It is expected that the SIR improvements for the other cases described in reference [3] suffer approximately the same amount of impairment. Therefore, the diversity combination of the present invention provides the wireless communication systems with a better SIR of at least 3 dB over conventional diversity selection of two antennas [reference 3].

Table 1 presents a summary of error statistics in the hardware simulations of the 1483 independent bursts received. The average value of WER is defined as the ratio between the number of bursts containing by at least one one-bit error and the total number of simulated bursts. The average value of the WER for a single antenna and selection in diversity is 17.6% and 9.1%, respectively. It is not possible to have reliable voice communications with these WER values so high. Note that the diversity selection WER can not be approximated by the WER square of a single antenna. This is because there is a correlation between the signals received in the two antennas caused by the normal logarithmic fading, even though the Rayleigh fading is independent between the antennas. With the combination in diversity, the average value of the WER is 4.9%, which is a marginally acceptable value for reliable voice communications. Using channel access procedures, based on power and signal quality measurements, the system performance can be improved by a total of 7.4 dB [reference 3]. This further improvement of the SIR will make possible more reliable voice communications under these conditions, by using the diversity combining process of the present invention.

Table 1: Computer hardware simulation results for 1483 received bursts. The handset 25 is located in the worst area from a statistical point of view. It incorporates macroscopic diversity based on the measurement of channel power. The values of Cs, n and s are 8, 4 and 10 dB, respectively.

Table 1 also shows that the diversity selection of two antennas retrieves 48.3% of the bursts that had been lost in the case of a single antenna without diversity. The diversity combination of two antennas of the present invention recovers an additional percentage of 45.9% of the bursts that had been lost in the diversity selection of two antennas. 4. 3 Frequency selective fading mitigation This section describes the studies of the effectiveness of diversity combining and diversity selection to combat the distribution of delay (selective frequency fading). Common channel interference is not included in these delay distribution simulations. In this study two power delay profiles are used, the "two-ray" profile (two rays) and the truncated Raylegh profile, which are shown in Figures 12A and 12B, respectively. In Figure 12A, T = 2DS, where DS stands for "Delay Spread" or delay distribution. In Figure 12B, the channel impulse response at time iT is a random Gaussian process with zero mean value and P variation. P i = (i / 4) e "i2 / 32 T2 = (DS) 2 / 4.7 Table 2 shows a summary of the hardware simulation results for the irreducible word error ratio (WERs) of a single antenna, the selection in diversity and the diversity combination of the present invention. All WERs, except 0.1% and 0.3% indicated in Table 2, contain more than 50 cumulative burst errors. Even for the limited number of accumulated errors (WERs of 0.1% and 0.3%), the WERs are probably less than 1%. The normalized delay distribution (d) is defined as the relationship of the delay distribution (DS) and the symbol period. WERuna so / a ßntm, WERselection and WERcombination are the irreducible relations of word errors for the cases of a single antenna, selection in diversity and combination in diversity of the present invention, respectively.

When d = 0.2, the WERs for both power delay profiles are approximately the same in each individual case (a single antenna, selection and combination). This is reasonable, since the operation of the system is not sensitive to the shape of the power delay profile when the normalized delay distribution is small [reference 10]. The WERs for a single antenna and the selection in diversity for d = 0.2 are reasonably well in accordance with the values presented in reference [8], which indicates WERs for values of "d" of up to 0.25.

Table 2: Average simulated irreducible WERs for the "two-ray" power delay profile (two rays) of Figure 12, with values of d = 0.2, 0.4, and the truncated Rayleigh power delay profile of the Figure 12B with values of d = 0.2, 0.3 and 0.4. All WERs, except the two cases of less than 1%, contain more than 50 cumulative burst errors. The diversity combination of the present invention produces a worse system performance for the truncated Rayleigh power delay profile of Figure 12B than for the "two-ray" power delay profile (two rays) of the Figure 12A. This is to be expected because, with the two antennas 10, 20, you can only suppress or optimize a radio path. With d = 0.3 in the truncated Rayleigh power delay profile model, diversity selection has an irreducible average WER of 2.4%, which can still provide relatively good voice communications. The diversity combining WER is expected to be smaller in the "two-ray" power delay profile model when d = 0.3. Of the studies presented in reference [8], for the same WER of 2.4%, the models of diversity selection and of a single antenna without diversity can sustain a normalized delay distribution of up to 0.16 and 0.06, respectively.

The above description is included to illustrate the operation of the preferred variations and is not intended to limit the scope of the invention. The scope of the invention will be limited only in accordance with the following claims. From the foregoing discussion, many variations will be apparent to a person skilled in the art and such variations are included in the spirit and scope of the invention.

Claims (19)

CLAIMS Having sufficiently described my invention, I consider as a novelty and therefore claim of my exclusive property, what is contained in the following claims ^ 1. A method for combining in diversity a first and a second electromagnetic signal within a receiver, characterized in that it consists of the following steps:
1. I. - coupling of two antennas to said receiver, a first antenna for receiving said first signal and a second antenna for receiving said second signal; and II.- coupling to at least one of the antennas of a circuit to vary the gain and phase of the signal received through said antenna, with said gain and phase constrained to be selected within a finite set of discrete gains pre-selected and a finite set of preselected discrete phases.
2. 2. The method of claim 1, wherein there are several transceivers each containing one of said receivers, with said transceivers located geographically so that said transceivers can communicate with each other using radio frequencies.
3. 3. The method of claim 2, wherein at least one of said transceivers is a portable hand-held transceiver.
4. 4. The method of claim 2, wherein at least one of said Transceivers is part of a base station.
5. 5. The method of claim 1, which also contains the scanning step through at least one of the gain and phase combinations to obtain the optimum signal-to-signal deterioration ratio within said receiver.
6. 6. The method of claim 5, which also contains the step of establishing gain and phase when said optimum signal-to-signal deterioration ratio is obtained.
7. 7. The method of claim 5, wherein the scanning step is divided into two discrete scans, a relatively rough scan and a relatively accurate scan.
8. 8. The method of claim 5, wherein the receiver receives the signals according to a time-division protocol, such that said signals are sent in bursts and said scanning step is executed once per burst.
9. 9. The method of claim 5, wherein said scanning step is executed each time the signal-to-signal deterioration ratio falls below a preselected value.
10. 10. The method of claim 1, wherein the receiver is part of a transceiver that contains a transmitter and in which said receiver and said transmitter operate on the same frequency, and similar steps executed by the receiver are also executed by the transmitter.
11. 11. Of an apparatus for the diversity combination of a first and a second signal, characterized in that said apparatus consists of: • a receiver; • two separate antennas coupled to said receiver, a first antenna coupled to said first signal and a second antenna coupled to said second signal; • an analog-to-digital converter coupled to each antenna; "A time delay circuit coupled to at least one anaerobic-to-digital converter consisting of a set of time delay elements adapted to insert a fixed amount of phase change; • a stepped amplifier / attenuator coupled to at least one analog-to-digital converter to insert a fixed amount of gain; • a signal combiner coupled to the time delay circuit (s) and to the amplifier (s) / attenuator (s); and • a circuit for measuring the signal-to-signal deterioration ratio within said receiver coupled to the signal combiner.
12. 12. The apparatus of claim 11 further comprising, coupled to said signal combiner, a signal-to-signal deterioration and symbol synchronization measurement circuit.
13. 13. The apparatus of claim 12 also containing at least two signal-to-signal deterioration and symbol synchronization signal circuits connected in parallel.
14. The apparatus of claim 11, wherein said receiver employs QPSK modulation and whose apparatus also contains: • a l / Q component generator coupled to said signal combiner; Y • a phase recovery circuit coupled to said component generator l / Q.
15. 15. The apparatus of claim 11, wherein said two antennas are separated by some combination of spatial separation, polarization and radiation pattern.
16. 16. An apparatus for the diversity combination of a first and a second electromagnetic signal, characterized in that said apparatus consists of: "a receiver; • two antennas coupled to said receiver, a first antenna for receiving said first signal and a second antenna for receiving said second signal; • coupled to at least one of the antennas, a circuit to vary the gain and phase of the signal received by said antenna, with said gain and phase constrained to be selected from a finite group of preselected discrete gains and a finite group of pre-selected discrete phases; and • coupled to the circuit that varies the gain and phase, a combiner to combine the two signals into a combined signal.
17. 17. The apparatus of claim 16, which also contains, coupled to the circuit that varies the gain and phase, a circuit for scanning through at least some of the gain and phase combinations to optimize the signal-to-speech ratio. deterioration of signal within said receiver.
18. 18. The apparatus of claim 17, which also contains, coupled to the scanner circuit, a circuit for adjusting the gain and phase once said signal-to-signal deterioration ratio has been optimized.
19. 19. The apparatus of claim 16, which also contains, coupled to the circuit that varies the gain and phase, an intelligent rounding circuit to recover the carrier phase of the combined signal without any significant impairment in system performance.