MX2008005275A - Iteractive frequency correction using training sequence and data bits - Google Patents

Abstract

A receiver for receiving transmitted signals comprising reference information known to said receiver, said receiver comprising means for estimating said reference information and non reference information to provide estimated information, said non reference information being at least one of immediately preceding and immediately succeeding said reference information;andmeans for using the estimated information to determine an error said received signals and to compensate for said error.

Description

INTERACTIVE FREQUENCY CORRECTION USING TRAINING SEQUENCE AND DATA BITS FIELD OF THE INVENTION The present invention relates to a receiver and a reception method. In particular, but not exclusively, the present invention relates to a method of receiving a signal containing information, depending on the frequency or some characteristic related to the frequency.

BACKGROUND OF THE INVENTION Wireless communication systems are known. A known system is illustrated schematically in Figure 1. The area covered by a wireless communication network (2) is divided into several cells (4). The cells can be side by side, overlapping, or both. Each cell (4) is provided with a base station (6). Each base station (6) is configured to communicate with mobile stations (8) or other user equipment located in the cells. Several different standards are known that regulate the communication between the base stations and the mobile stations. A standard commonly used is the standard of the Global System for Mobile Communications (GSM, for its acronym in English). This is a digital communication system. In GSM, the data is transferred between the mobile stations (8) and the base stations (6) as a radio signal on a physical channel that can use time or frequency division multiplexing or both to create a sequence of radio frequency channels and time intervals. Each frequency band is divided into multiple access frames by time division, with 8 users per frame. Each user is given time to send a simple burst of information. Commonly, the mobile station and the base station that are in communication will use different frequency bands. In some implementations, GSM can use Gaussian Minimum Displacement Transmission (GMSK) modulation. The GMSK modulation uses the phase of the radio signal to transmit the data. Obviously, the phase of the signal depends on the frequency of the signal. In order to correctly identify the transmitted data, the frequency of the signal received at either the receiving base station or the mobile receiving station must be within defined limits, as compared to the desired transmission frequency for that signal. If the frequency has shifted beyond these limits, then errors in data recovery may occur.

Errors may occur in the frequency, in one of the receiving mobile station and the receiving base station, for various reasons. For example, this may occur if the mobile station or the base station is in motion. Generally, it is clear that the mobile station will move. Obviously, changes in frequency due to Doppler shift may occur. This effect is particularly noticeable when the mobile station is moving relatively fast. For example, high-speed trains are being proposed, with a speed of approximately 330 km / hour. At those speeds, the Doppler shift introduced by the movement of the mobile station would result in a relatively large frequency change. Of course, it should also be understood that moving at slower speeds will also result in Doppler shifts. The movement of the mobile station in relation to the base station is not the only source of frequency changes. Other errors may occur. For example, multipath propagation may change the frequency of the received signal. It is possible that the transmitter oscillator is not working properly, for example, due to changes in temperature and consequently, the transmitted signal and hence the received signal would not be at the correct frequency. Also, adverse weather conditions, particularly a very hot or very cold climate, they could change the condition of the radio channel, which again would result in a frequency shift of the received radio signals. In general, changes in frequency occur either due to deterioration of the radiofrequency or changes in the characteristics of the channel. The deteriorations of the radio frequency include multipath propagation and variation in the oscillating characteristics of the crystal. The change in channel characteristics include effects due to movement and changes in weather conditions. Generally, the GSM standard is reasonably reliable. As such, it has the ability to adapt to some variation in frequency. However, there may be errors coming from more than one source, which generate, cumulatively, a relatively large frequency error. Also, mobile stations that move very fast can cause, by themselves, a relatively large frequency shift. Reference is made to International Patent Application WO 03/039025 in the name of the present applicant. In this document, the automatic frequency correction is described. In a first stage, the frequency is estimated with the use of a training sequence portion. Subsequently, the estimated frequency compensation is removed from the samples and taps. In a second step, some symbols are preequalized using a decision feedback equalizer. Frequency compensation is estimated with the use of the training sequence, queues and extended symbols. Then, the frequency compensation of the samples and shots is eliminated. This configuration has the potential problem that the performance of the global automatic frequency correction depends on the first stage. However, the first stage only uses the training sequence. The frequency compensation estimate with the use of the training sequence portion alone can not be reliable when the signal-to-noise ratio is reduced. In this context, the decision feedback equalizer can introduce more errors than the decisions taken without the first stage and, in this way, affect the entire performance. In addition, fast synthesizers are being considered, which can jump between time slots. However, this implies several restrictions on the algorithms of the digital signal processor. A tail and a few symbols may be corrupted and unusable. Due to the proposed jump between time intervals, the stabilization time of these synthesizers is a cost function and the stabilization time may, for example, be of the order of 20-30 microseconds. As a result of this stabilization time, some of the symbols and tails become unusable.

SUMMARY OF THE INVENTION According to one aspect of the invention, there is provided a receiver for receiving transmitted signals comprising reference information known by the receiver; The aforementioned receiver comprises means for estimating said reference information and non-reference information, to provide estimated information; the information that is not of reference is the information that immediately precedes or immediately follows (or both) the aforementioned reference information; and means for using the estimated information to determine an error in the received signals mentioned and to compensate for said error. According to another aspect of the invention, a receiver of transmitted signals comprising reference information known by the receiver is provided; the aforementioned recipient comprises an estimator for estimating the mentioned reference information and the non-reference information for providing estimated information, the non-reference information is the immediately preceding information or the immediately following information (or both) to the reference information mentioned; an error estimator to use the estimated information to determine a frequency error for the received signals; and a frequency error corrector to compensate for the error of the received signal. In accordance with another aspect of the invention, there is provided a method of receiving transmitted signals comprising known reference information; the method of reception includes, as steps, estimating the aforementioned reference information and non-reference information to provide estimated information, the non-reference information is the information that immediately precedes or immediately follows (or both ) to the reference information mentioned; and use the estimated information to determine an error in the received signals mentioned and to compensate for said error. According to another aspect of the invention, there is provided a method of frequency correction comprising a first stage and a second stage, the first stage comprising a first means for estimating a first number of symbols of a received signal; a first means for estimating the frequency error based on the estimated symbols and a first means for effecting a frequency compensation correction on a second number of symbols of the received signal, to provide a modified signal, while the second stage comprises a second means for estimating at least the second number of symbols of the modified signal, a second means for estimating the frequency error based on the emission of symbols estimated by the second estimation means and a second means for effecting a correction of frequency compensation in the received signal with the use of the estimated frequency error by the second estimation means.

BRIEF DESCRIPTION OF THE DRAWINGS 0 FIGURES For a better understanding of the present invention and to know how this invention can be put into practice, reference will now be made, by way of example, to the accompanying drawings, in which: 1 shows a schematic view of a network; Figure 2 shows a schematic representation of a burst in the GSM standard; Figure 3 shows the general structure of a receiver that can be used in embodiments of the present invention; Figure 4 shows part of the receiver of Figure 3 in greater detail; Figure 5 shows a modified burst structure; Figure 6 shows a flow chart of the steps that are followed in the embodiments of the present invention; and Figure 7 shows a second embodiment of the present invention.

DETAILED DESCRIPTION OF THE MODALITIES OF THE PRESENT INVENTION Reference will now be made to Figure 2, which shows a schematic representation of a GSM burst. In the GSM standard, the data modulated with GMSK or with 8PSK are generated in a burst containing a sequence of 156.25 complex symbols. Each symbol has a real part and an imaginary part. 156.25 is the number of symbols that can fit in a time interval. The burst (22) has six components. A first "bit-tail" field (26) is provided. This first tail bit field (26) comprises three symbols. Next is a first group of coded data (28). This consists of 58 symbols. This is followed by a training sequence (24) which is 26 symbols only. The receiver knows in advance the training sequence. This is known as a midamble sequence, since it comes between two data fields. The training sequence (24) is followed by a second data group (30), which again comprises 58 encoded data symbols. Finally, this is followed by a second "bit-tail" field comprising 3 symbols. At the end of the burst is a protection period (34), which is empty and extends for a period equivalent to 8.25 symbols. In the GMSK modulation scheme, the symbol is equivalent to one bit, so there are 148 bits in a burst. It should be appreciated that the receiver knows in advance the symbols of the training sequence. Normally, the frequency of the burst received by either a receiving mobile station or a receiving base station varies with respect to the desired transmission frequency for the burst by a certain amount of "frequency compensation". This frequency compensation has the consequence of the phase of the signal change with time. The embodiments of the present invention are configured to compensate some of the frequency compensations, regardless of the cause. For example, frequency compensation may occur due to movement of the mobile station, changes in temperature, changes in the characteristics of the components or the like.

Reference will now be made to Figure 3, which illustrates schematically a receiver exemplifying the present invention. The receiver may be incorporated in a mobile station, a base station or both. The transmitted bursts (22) are received in an antenna (12). This signal received by the antenna (12) is introduced into an amplifier (14) that amplifies that signal. The amplifier sends the amplified signal to the filters and mixers (16). The filters and mixers (16) eliminate the signals that fall outside the frequency of interest and also reduce the signal to the frequency of the baseband. The emission of the filters and mixers (16) is introduced in an analog to digital converter (18). The emission of the filters and mixers is an analog signal. The analog to digital converter (18) converts this analog signal into a digital signal. The digital signal is input to a digital signal processor (20). Preferably, the embodiments of the present invention are carried out in a digital signal processor or at least in a digital environment. However, it is conceivable that the alternative embodiments of the present invention can be carried out, at least partially, in the analogical field. Reference is now made to Figure 4, which shows part of the receiver of Figure 3, in greater detail. The first block, Reduced State Sequence Estimation (RSSE) (102) receives the training sequence by the training sequence block (100). This is a local version of the training sequence and not the version of the training sequence that has been transmitted. The RSSE block (102) also receives a channel impulse response input h and samples and received. This block estimates 10 symbols on each side of the training sequence. In other words, the equalization of 10 symbols of the received data is estimated, placed on either side of the training sequence. The 20 estimated symbols, 10 on each side of the training sequence, XRSSE and the 26 known training symbols (XTRS) are introduced to a frequency estimator block (104) of Linear Mean Square Error (LMMSE). English) . This block (104) also receives as an input, the channel impulse response h and the samples received and. The LMMSE frequency estimator block (104) uses the 10 estimated symbols on each side of the training sequence as well as the 26 training sequence symbols to estimate the frequency compensation. In this way, the output of the LMMSE frequency estimating block (104) is fDopierTp which is the calculation of the Doppler frequency compensation, which is input to the frequency compensation correction block (106). In addition, the frequency compensation correction block (106) receives the channel impulse response h and the samples received and. The estimated frequency offset is corrected for the samples and the received sockets. However, this correction is only carried out for the training sequence and 30 data symbols on each side of the training sequence. The frequency correction block (106) transmits the modified channel pulse response A and the corrected received samples and? to the Decision Feedback Equalizer / Reduced State Sequence Estimation block (DFE / RSSE) (108). In addition, this block (108) receives the training sequence of the block (100). Block (108) is configured to equalize, ie estimate, 30 symbols on either side of the training sequence. In other words, a total of 60 symbols are estimated. This block uses the modified channel impulse response and the corrected samples. The block (108) transmits the extended extended symbols XExtene to a second LMMSE frequency estimating block (110). The second frequency estimating block (110) also receives the impulse response of channel h and the samples received and, as well as the training sequence. Then, the frequency is estimated with the use of the 60 preequalized or estimated symbols and the 26 symbols of the training sequence. Subsequently, the second frequency estimator provides a second estimate of the frequency Doppler füoppierFinai that is input to a second frequency compensation correction block (112). The second frequency compensation correction block also receives the original channel impulse response h and the samples received and. The second frequency compensation correction provides a modified channel impulse response h "and modified samples and" to the second stage, the final stage of frequency compensation correction. Thus, unlike the prior art, the final frequency compensation correction applies the final correction in samples and shots in which the frequency compensation has not been corrected. It has been found that, by applying frequency compensation correction in the original samples, better performance is provided than if the correction is applied in the corrected intermediate samples. In this way, the embodiments of the present invention have the ability to improve the performance of AFC driven by decision by improving the decisions of the preequalizer. This will apply both for the automatic frequency correction directed by decision based on queue as well as without queue. It has been found that this can address the improved performance problems of the base station, which require that the stabilization time take into account the frequency jump. At least some of the embodiments of the present invention have the ability to approach enough to match a performance of the known algorithm used by the queue. In some embodiments of the present invention, the tail symbols may be used additionally. The EDGE (Enhanced Data Transfer Rate for GSM Evolution) technology is considering introducing a new normal burst format to improve the efficiency of mobile multiple-interval classes. One of the options that is being considered to increase the payload data transfer speed is to eliminate some or all of the predetermined data from the time interval after the initial interval, which is for example, the tail bit, FACCH tags, training sequence, and / or protection period of the assigned multiple intervals 2 .... n. To maintain channel tracking capabilities, frequency compensation may need to be updated for the assigned ranges 2 ... n. In this case, it is necessary to generate the reference data for frequency estimation, with the use of the same principle DFE / RRSE as it is done for the extension of the training sequence, as established in relation to the modalities of the present invention . Reference is made to Figure 5, which shows a proposal for a possibility of multiple intervals of phase 2 of EDGE. In this example, the multiple interval is extended to cover two intervals. In the proposals, there is a first field (126) comprising queues. This is followed by a first data field (128). An indicator field (146) is provided which contains an indicator symbol. This is followed by the training sequence field (124). A second indicator field (148) is provided which follows the training sequence field. It should be taken into account that the indicator fields can be provided in the modality shown in figure 2. Until now, the structure of the field is similar to that shown in figure 2. However, this is followed by a field of data that extends from one interval to the next interval and contains many more symbols. For example, the current proposal for the second data interval contains 57 symbols. The proposal illustrated in figure 5 has 213.25 symbols. This extended data interval is followed by an interval queue (132). It must be taken into account that the length of the second data field can be extended to cover even more intervals. In this way, the elimination of the training sequence of the GSM burst format frees up more space for additional user data. Thus, within a multiple interval assignment, this could be done from the interval 2 .... n, that is, the training sequence that remains in the first interval and would disappear in the next. The indicators can be eliminated from the interval 2 .... n, since they would apply the indicators of the interval. The embodiments of the present invention can be applied particularly in the slot structure shown in FIG. 5. The embodiments of the present invention, two taps, two states that reduce the complexity equalizers, are used for pre-equalization. With this type of equalizer, the complexity to estimate 20 hard symbols is relatively insignificant. In a preferred embodiment of the present invention, it has been suggested that in the first estimate 10 symbols are used on either side of the training sequence. In the first frequency estimate, the 20 preequalized symbols and the 26 symbols of the training sequence are used, resulting in 46 symbols. In the second frequency estimator, 30 symbols are used on either side of the training sequence. It should be taken into account that the number of symbols used may vary in some alternative embodiments of the present invention. For example, in the first equalization block, the number of symbols on either side of the training sequence could preferably be between 10 and 15. Preferably, the number of symbols used by the second equalizer is approximately the same as the number of symbols. training sequence symbols. The processing part of the digital signal processor can effectively define the upper limit as the number of symbols that are processed on either side of the training sequence. In preferred embodiments of the present invention, the symbols on either side of the training sequence are equalized. However, in alternative embodiments of the present invention, only the symbols on one side of the training sequence can be used. The symbols can be directly adjacent to the training sequence or be eliminated from it by a small number of symbols.

Reference is now made to Figure 6, which shows the steps of the method exemplifying the present invention. In step (YES), the 10 symbols are equalized (estimated) on either side of the training sequence. c In step (S2), the Doppler frequency compensation is estimated using the 10 symbols on either side of the training sequence and the training sequence. In step (S3), the estimated Doppler frequency compensation is used to correct 30 symbols on either side of the training sequence as well as the received training sequence. The channel impulse response corrected by frequency and the received samples are used in step (S4). In particular, the 30 symbols are equalized (estimated) on either side of the training sequence. This is based on the originally received symbols and the channel impulse response initially determined. The estimated 30 symbols on either side of the training sequence are used in step (S5) along with the training sequence to estimate the frequency. The second estimate of the Doppler frequency is calculated based on the channel impulse response and the samples as received, without any correction. In other words, in step (S5) the same version of the channel impulse response and the received samples as in step (SI) are used. The estimated frequency is used in step (S6) to correct the channel impulse response initially determined and the samples received. Reference is now made to Figure 7, which shows a second embodiment of the present invention. In this mode, the average is used. The average is necessary to achieve a high-speed requirement. Due to the increased speed of training, it has been desired to improve the performance of the base transceiver station to maintain the same quality requirement, particularly for GMSK modulated signals. The base transceiver station will meet must meet an average quality level 4 or higher to achieve a successful process of preparation and transfer with mobile devices that move at a speed of up to 330 km per hour. Blocks (100), (102), (104), (106), (108), (110) and (112) operate in a manner similar to that described in relation to Figure 4. However, instead of receive the channel impulse response h and received samples and, the blocks (102), (104), (106), (110) and (112) receive the modified channel impulse response Ji 'and the output of modified received samples and 'by means of the average elimination block (180). An additional minimum squared error (LSE) frequency estimator (184) is provided. This receives the output from the block (108) as well as the output from the average elimination block (180). This provides an estimate of the frequency compensation using the least squared error frequency algorithm. The block (184) also uses the 30 symbols on either side of the training sequence as well as the training sequence itself. The frequency compensation calculated by the frequency estimating block LSE (184) is sent to the average update block (182). The latter provides an average of frequency compensation. Then the frequency compensation averaged in the average elimination block (180) is used. The embodiments of the present invention can be incorporated into a base station and / or a mobile station or other suitable user equipment. The preferred embodiment of the present invention has been described in the context of the GSM system using GMSK modulation. It should be appreciated that the embodiments of the present invention can be used with different modulation methods, which are based on frequency or on a frequency-dependent characteristic. Obviously, the embodiments of the present invention can be used with any other communication method or standard with the modulation used, which depends at least on the frequency. The embodiments of the present invention only apply to wireless cellular communication systems, but can be used in any configuration where the signals are transferred using a modulated radio signal or the like.

Claims (21)

1. CLAIMS: 1. A receiver for receiving transmitted signals comprising known reference information by the receiver, wherein said receiver comprises: means for estimating the reference information and non-reference information for providing estimated information; the information that is not of reference is the one immediately preceding or the one immediately following (or both) the aforementioned reference information; and means to use the estimated information, to determine an error in the received signals and to compensate for the error.
2. 2. A receiver according to claim 1, wherein the non-reference information comprises n symbols that precede or follow (or both) the reference information. A receiver according to claim 2, wherein n is between 8 and 15. 4. A receiver according to any of the preceding claims, wherein the reference information comprises a training sequence. 5. A receiver according to any of the preceding claims, wherein the reference information is provided in an intermediate region of a signal. 6. A receiver according to any of the preceding claims, wherein the estimating means comprises an equalizer. A receiver according to any one of the preceding claims, wherein the estimating means is configured to carry out one of a reduced state sequence estimation method and a decision feedback estimation method. 8. A receiver according to any of the preceding claims, wherein the error comprises a frequency error. 9. A receiver according to any of the preceding claims, wherein the means for using the estimated information comprises a frequency estimator. 10. A receiver according to claim 9, wherein the frequency estimator uses an LMMSE method to estimate the frequency. A receiver according to any of the preceding claims, wherein the means for using the estimated information is configured to provide a frequency compensation correction for at least a part of a received signal. A receiver according to any of the preceding claims, wherein a second means for estimating reference information and non-reference information is provided to provide estimated information; the second estimation means is connected to receive the emission of said means to use the estimated information. 12. A receiver according to claim 2 and 11, wherein the second estimation means is configured to estimate in symbols of non-reference information, which precedes or follows (or both) the reference information, where m is greater than n. 13. A receiver according to claim 13, wherein m is in the range of 25 to 35 symbols and preferably 30 symbols. A receiver according to claim 12 or 13, wherein the second means for estimation is configured to estimate only the m symbols of the non-reference information, which precedes or follows (or both) the reference information. 15. A receiver according to any of claims 11 to 14, wherein a second means for using the estimated information of the second estimation means is provided. 16. A receiver according to claim 15, wherein the second means of use is configured to estimate a frequency compensation using the received signal. 17. A receiver according to claim 15 or 16, wherein the second means of use comprises a means for correction of frequency compensation, to correct the frequency of the received signal. 18. A receiver for receiving transmitted signals comprising reference information known by the receiver; The receiver includes: an estimator to estimate the reference information and non-reference information to provide estimated information, the non-reference information is the information immediately preceding or immediately following (or both) the reference information mentioned; an error estimator to use the estimated information to determine a frequency error for the received signals; and a frequency error corrector to compensate for the error of the received signal. 19. A method of receiving transmitted signals comprising known reference information, the receiving method comprises the steps of: estimating reference information and non-reference information to provide estimated information, non-reference information is the information that immediately precedes or immediately follows (or both) the aforementioned reference information; and use the estimated information to determine an error in the received signals and to compensate for the error. 20. A method of frequency correction comprising a first stage and a second stage, the first stage comprises a first means for estimating a first number of symbols of a received signal, a first means for estimating the frequency error based on the estimated symbols, and a first means for carrying out a frequency compensation correction in a second number of symbols of the received signal, to provide a modified signal, and the second stage comprises a second means for estimating at least the second number of symbols of the modified signal, a second means for estimating the frequency error based on the output of symbols estimated by the second estimation means, and a second means for carrying out a frequency compensation correction on the received signal using the frequency error estimated by the second estimation means. 21. A receiver according to claim 20, wherein the first number is smaller than the second number.