MXPA00001778A - Indoor communication system and synchronisation for a receiver - Google Patents

Abstract

In a multi-user system for wireless transmission of video, audio and/or combined signals or general communication signals a proposal is given for the timing and synchronisation of the various messages of the users. The proposal is based on a combined FDMA/TDMA transmission and access technique. Data from the users are transmitted in a common frame comprising a number of control slots and also a number of data slots, each slot being associated with a guard time. The slot signals have the characteristics that i) the time position of the first control slot signal in the frame is based on the last control slot signal of the previous frame, ii) the time position of the subsequently transmitted control slot signals is based on the position of the previous control slot signal, iii) the time position of the data slot signals is based on or is referred to the last control slot signal in the same frame, which can be a signal of another user or of its own, iv) at the end of the frame, behind the last slot signal an additional or increased guard time significantly longer than the general guard times associated with the other slots is placed. The proposal also includes a communication system and a receiver for a communication system.

Description

INTERNAL COMMUNICATION SYSTEM AND SYNCHRONIZATION FOR A RECEIVER For private households and also for local area networks (LAN) developers are looking to connect all kinds of devices such as television, personal computer, stereo system, alarm system, telephone, etc., together. Household systems that communicate by the use of the 230 volt power line are already known. The object of the invention is to provide a time reference system and a synchronization method for a receiver of the internal communication system, above. All devices in a home have to work in an almost-synchronous mode with a frequency stability, for example, of approximately 10 x 10"6. The transmitted, inventive signals of any device are aligned in frame to the signals of the others The synchronization of the following frame is based, according to the invention, directly on the signal of the last preceding control segment The signals of the control segments, transmitted are related to the control segment, precedent, respective, the The first control segment of the table is related to the last control segment of the previous frame. The synchronization of the data segment signals of all the devices is based on the signal of the last control segment of the same frame, which applies the defined frame structure. At the end of the table a longer safety time can be included. A receiver according to the invention is described for frequency synchronization of a device. To synchronize different signals within a frame that are incorrect in their relative timing and differ from their RF frequency more or less, an ad-hoc process is necessary. From the correlation of the half-path, a first correlation signal is derived. With this first correlation signal, a channel compensation process is initiated which produces a corrected output signal of the channel. From this corrected output signal of the channel, the signal can be reconstructed identically to the input signal by using the channel impulse response. 1 . REQUIREMENTS FOR THE S INCONNECTION OF THE SI STEMA - Pe inrt e s for the s e n a grouping, and in some aspect also for the identification of signs of other groupings. 1. 1 INTRODUCTION The various signals of the users and terminals of a cluster are embedded in a common TDMA frame. If the use of a second channel is allowed, this embedding applies separately for both channels, while the messages and the transmission of data of a user can be divided over the two channels. The organization of tables, inventiveness is based on time segments that are assigned to the users during the initialization processes and which requires at least an almost synchronous generation in time and a transmission of the segment signals. Additionally, certain synchronization tolerances have to be maintained in order to avoid collisions within the frame itself and to ensure the transmission of the defined data cups. In addition, the evaluation of the various signals in the receiver should be as simple as possible, especially where the maximum deviation of the center frequency could be an important factor, for example, for the correlation with the means- tours and the channel compensation process. The principles are described and the tolerance proposals are given in chapters 1.2 and 1.3. In particular, the tolerances can be subject to additional considerations. The transmission, especially the 'correct positioning of the segment signals, transmitted, requires a suitable receiver in order to monitor and evaluate the signals of the other users over time and, consequently, fix the signals themselves. The corresponding part is described in chapters 2, "receiver synchronization" and 3"acquisition processes, monitors and channel perception". Since the various signals of the picture are not completely synchronous, neither in the RF frequency nor in the synchronization; The receiver has to deal with changing conditions and an almost ad-hoc synchronization or evaluation must be performed if the signals from more than one user are inspected or evaluated. 1. 2 RF ACCURACY (FREQUENCIES) Depending on the requirements, the reference oscillator (s) can be a factor of remarkable cost. On the other hand, a very complex processing in the receiver, in order to deal with rather bad conditions, could also cause realization problems. Finally, a commitment has to be made. A requirement from the point of view of the receiver could be to allow a quick monitoring with a practicable means, that is, with only one correlation per segment or half-way, which implies that the phase rotation within the half-path caused by the frequency deviation is significantly less than 180 °; otherwise the correlations with pre-distorted half-courses have to be made or the result in critical cases will be significantly less than the maximum. Taking into account the relatively high frequencies of the ISM bands with 2.4 and 5.7 GHz and additionally, the fact that the receiver has worked with the corresponding deviations of the transmitted signal and its own oscillators, this leads to a tolerance of + _ 10 * 10 ~ 6 The corresponding frequency and phase deviations are shown in table 1.1.

TABLE 1.1: FREQUENCY (RF) AND PHASE DEVIATIONS BASED ON A RELATIVE TOLERANCE OF ± 10 * 10"6 Frequency deviation for example + 24kHz corresponds to a relative "phase velocity" of 24,000"signal rotations" (2p) per second equal + 24,000 / s * TS * 360 ° approximately + 1.05 ° per symbol.

Deviations on a few symbols are already too large to be ignored in the receiver. The methods of evaluation of impulse and search response, maximum, dedicated and also correction methods and special signals or equivalent measures are needed in order to achieve sufficient results in the correlation and demodulation processes. On the other hand, a crystal oscillator with an accuracy and long-term stability of + 10 ppm is already a sophisticated device for consumer applications and seems not to be adequate to place more severe restrictions on this part of the system. 1. 3 SYNCHRONIZATION OF TABLE AND SYNCHRONIZATION OR TIMING OF SEGMENTS AND SYMBOLS The transmitted signal (s) of any user must be aligned in frame to the signals of the others (if present), which implies: Place and number the signal of the control segment itself according to the placement and numbering of the control segment signals of the other users (preferably, the one that is at the top in the series of signals). In addition, to concatenate all the signals of grouping and avoid the formation of sub-groupings, the synchronization of any control segment signal must be based directly on the preceding signal of the control segment, transmitted by another user, or, if there is no other user, by its own transmitter in the previous frame; normally a signal within the same frame, but in the case of the first, it must be the signal of the last control segment from the previous frame. 20 The synchronization of the data segment signals of all users will be based on the signal of the last control segment having a relevant amplitude, applying the defined frame structure.

It should be noted that under certain conditions, acquisition phase, the control signals should not be transmitted in all frames; The principles are described in the system description. This requires a flexible operation of the device with respect to the segment signal to be used as a reference. Further details of the preferred embodiments of the invention are described with reference to the figures.

Figure 1: Alignment tolerances and synchronization of signal frames within a frame (example), Figure 2: concatenation of several signals and tolerances Figure 3: Concatenation of signals and tolerances (b and c); the choice of erroneous reference signals.

Figure 4: An example for a correlation result (only magnitudes), comprising the components of three users (different magnitudes and constellations of delays).

Figure 5: Correlation result showing the signals of Figure 4 after assignment to a commonly used detected frame (partially different time separation, noise or suppressed interference).

Figure 6: Functional diagram of channel estimation, compensation and frequency correction (phase).

Figure 7: Correlation result containing control and data signals from three transmitters (separation during partially different time, the threshold defines whether a segment is determined to be empty or not).

The alignment and tolerances of tables are shown in figure 1, Where: Ts ?? (= 25.8μs) and Ts? 2 (= 103.2μs) are the control durations and the theoretical data segment, including the safety time and Tfr (= 10.32 ms) is the theoretical duration of the table; nx and ny are the numbers of the complete spaced of the control segments between the segment taken into account and the reference segment (pertinent only within the section of the control signal) and m is' the number of the complete spacing of the data segments between the segment taken into account and the reference segment (relevant only within the section of the data signal); and ttoii, ttoi2 and ttoi3 are the permitted tolerances defined later. The components 0.558 Tsn and 0.486 Tsi2 take into account the duration of half control segment and half data segment, the first with a complete safety time and the second without a safety time, at the end of the section of the control segment. These values correspond to the definition of the center used later; they have to be adapted if another definition must be used. The following tolerances have to be applied in order to avoid deviations, greater synchronization over the length of a table and severe disadvantages in the handling of the data :.

TABLE 1.2 Segment Center of Any Control) *; with respect to the center of the tton = - 1.0 ... - 0.5 signal of the last segment of μs control received (from another user) *; 0 depending on the composition ttoi2 = -1.25 ...- 0.25 μs contemplated, whether t or ?? (with the same table) or ttoi2 (overlapping frames) Data segment centers) * ttoia = -1.0 ...- 0.5 μs with respect to the center of the + 25 * 10-d [(ny signal of the last segment of + 0.5Tsn control (derived from any + (ríiy + 0.5) Ta? 2] another user or that is a signal of itself) * 1) The centers of the transmitted signals of segment and the signal (s) of the reference control segments are defined herein by the centers of the half-way sections, placed between the tenth and the eleventh symbols, where in the case of a received reference signal, the first component of the response of impulse, with relevant amplitude, should be used. The time of Correlation and other delays caused in the receiver have to be removed. * 2) The placement is free if there are no other users, but in this case the tolerance ttoi2 has to be maintained to the signal itself in the previous frame.

Table 1.2: The synchronization tolerances have to be applied to the control and data segments. The terms used in the table are identical to those used in figure 1.

A negative tolerance ttoi ?, tto? 2 or tto? 3 means that part of the safety time of the preceding segment signal could already be used by the next signal. The reason is to ensure that the practical table can not be longer than the 10.32 ms defined, to guarantee the specified data rates. The processes will be based on the correlation of the media-routes transmitted, and in order to find the beginning of the frame, in the encounter of the signal (s) of the control segments. There are several possibilities; you can detect the frame by one or a combination of the following methods: First, the relative placement and sizes of the various segment signals can be evaluated, different size and spacing of the control segments and the general data segments; the 'method in some cases can not give a definite result. Second, a correlation with the sequence used for the identification number of the cluster, or a selected section of it, could be performed, which is a rather complex method but presumably produces a definitive result. • Finally, the content of all signals received from the segment (or only those that have been identified as being in the section of the control segment) will be analyzed with appropriate methods in order to detect the location numbers of the control segments, etc.

Additional details are described in chapter 2"Receiver synchronization". The solution requires an ad-hoc placement of the control segment signal and also commonly for all the data segments of a user, of the signals of the data segments. The tolerance of approximately 0.5 μs (approximately 4 symbols) allows the user to apply their own free run clock signal with + _ 25 * 10"6 and allows reasonable synchronization deviations in the evaluation of the segment reference signal and in signal synchronization for transmission The data segments are related to the last control segment in the frame and use the same clock along with the calculated and corrected distances Other synchronization solutions are possible, for example, based on any signal from segment in the preceding signal or based on all the signals of segment commonly in one, for example, the signal of the first control segment.The various requirements have to be satisfied in order to avoid the collision between the successive frames or sections of data within The following calculations confirm that this will be achieved. The maximum time, positive as negative by frame can be estimate by the following equation -? f r, (1 ... 16) ttoii + ttr, + tb Tf r where the part tto ?? represents the 16 tolerances (possible) in the signal placement of the control segment (including one of the previous frame) and 16 * ttr takes into account the sum of the transfer times (including an uncertainty of + _ 1/2 symbol) about the connections necessary to derive the reference position. Tb Tfr = + _ 25 * 10"6 * Tfr takes into account the basic (relative) tolerance of the reference system measured on a square (some of the distances to which these have been applied are shorter to some degree; negligible.) The corresponding internal distance T?, int is defined by the following equation: t?, mt = t Tfr + tto ?? 0 (condit.2) + ttr (condit 1) The application of the transfer time ttr depends on the question of whether the reference signal is derived from another user (condit 1) or the control signal itself (condit 2) if the signal is the last in the box).

The concatenation of both intervals with maximum and minimum values produces the time deviation tx (see below). Figure 2 shows the concatenation of the various signals and tolerances, where Xi ... X4 are the theoretical distances (see figure 1). tb is the basic tolerance (relative) of ± * 25 * 10 ~ 6, and tx represents a resultant deviation between the total and the internal structuring.

The requirements are: In order to achieve the defined data rate: (t?) Raax < 0; In order to avoid collisions between successive tables: [tb Tfr + ttoll + ttr.min + [tb fr + ttoll + ttr] max < 0.7 Tg; with Tg «3 μs, safety time. Another crucial issue may be that a user chooses the erroneous control segment signal (s) as reference as the placement of their own signals from the control and / or data segments, may be due to unsatisfactory reception conditions for the correct signal. In this case, the constellations according to Figure 3 are given and the resulting deviations t can be calculated by accumulating the differences of the minimum and maximum values according to the following formulas: for the control segments • y, c (max) s- [(ne + 1) (ttoll + ttr] min + [ne (ttoll + ttr)] < 0.7 Ta; and for the data segments y, d (maxx)) * - [tt Ti + t toll + t, ne (t toll + - tr J min + [tt Tfr + ttoll + ttrlmax < 0.7 Tg, - = - [tb Tfr + (ne + 1) (ttoll + ttrHmin + [tb Tfr + ttoll + ttr] max <0.7 Tg; where ne represents the number of error steps with respect to the correct reference position (for example with ne = 2, the reference signals correct and the next possible have been left out, they are not useful). The reference value is again 70% of the safety time. For control signals this is a hypothetical case more because only if the correct signal is detected, but not used (due to bad conditions), a corresponding position can be chosen (if the signal is not detected, the user can choose the same position that causes a collision). Figure 3 shows the concatenation of signals and tolerances (b and c); choosing erroneous reference signals. In all cases, minimum values can be calculated by changing the min / max conditions and average values are derived by averaging minimum and maximum results. With the tolerances defined and assuming a transfer time interval of -tr í0.01 ... 0.1 μs + 0.06 μs = -0.05 ... O.lß μs (equivalent to 2 ... 20 m and including. + _ 1/2 of uncertainty symbol), the results presented in table 1.4 are achieved: TABLE 1.3: MINIMUM, MAXIMUM AND AVERAGE DEVIATIONS WITHIN THE TABLE Due to averaging over a number of different effects and deviations of up to 16 users, the deviation intervals will presumably be smaller. In any case, a certain average deviation in the order of -8 μs per frame will remain, which means that the table is practically shorter than the defined value of 10.32 ms and the speed of possible data is greater by approximately 0.08%. Consequently, some of the occupied data sections, for example, a few symbols per segment from time to time, can not be filled and data management is needed in order to adapt the cups. Any kind of adaptation will be necessary in any case due to the wrong synchrony between the transmission medium and the data sources themselves. With respect to the choice of reference signals, erroneous, practically two steps are allowed for the control segments (83% with ne = 2), but only one step can be accepted for the placement of the data segments (75% with ne = 1) • In this way, the defined tolerance concept provides reasonable tolerances for system components and allows requirements to be met with respect to system performance, in order to guarantee specified data rates and / or avoid collisions. A preferred improvement of this solution is described by the following: If the safety times of all the segments are decreased by approximately 5% and the resulting time of approximately 15 μs will be added at the end of the This will allow it to significantly increase the tolerances. The values themselves move more to the center. This measure reduces the deviation of the possible average data cups. A "possible scenario" is shown below.

Second proposal Any control center segment) *; with respect to the center of t? = -0.5 ... 0.4 μs or the signal of the last received control segment (from another user) *; depending on the position ttoi2 = 0.75 ... 0.65 μs contemplated, whether tto ?? (within the same frame) or ttoi2 (frame overlap) Cent ro s of s tmentments of ttoi3 = 0. 5 . . . 0 4 μs dat or s) * with respect to the center of ± 25 * 10-6 * [(ny the signal of the last +0.5) Tsn control segment in the + (my + 0.5) Tsl2] box (derived from any other user or that is a sign of itself) Table 1.4 of chapter 1 (modified version). The synchronization tolerances to be applied to the control and data segments, modified values based on a small decrease of all safety times («5%) and an additional / longer safety time at the end of the frame. * additional details see Table 1.2 Results achieved with Table 1.5: minimum, maximum and average time deviations within the table (modified version, see above).

Therefore, with this mode, the safety time of the control segments is decreased to 2.75 μs, the safety time of the data segments is 2.85 μs, and the time gain of 16.48 μs is added to the end of the table, which produces 19.33 μs for the last value. 2. SYNCHRONIZATION OF THE RECEIVER 2.1 INTRODUCTION -see also chapter 1"Requirements for system synchronization". The frequency conversion and the demodulation and evaluation of the coherent signals in the receiver requires proper mixing and clock signals. In principle, if a continuous reception of only one signal or another device is taken into account, the reference oscillators for the frequency conversion and the clock could be synchronized in the received signals, or alternatively, a signal correction could be made received and downconverted and / or a correction of a free run synchronization oscillator signal. These processes could be based on the evaluation of the routes and the determined sections of the signals received from segments. However, in order to be able to synchronize rapidly different signals within a frame, which are incorrect in their relative synchronization and differ in frequency (RF), the synchronization has to be preferably an ad-hoc process instead of a continuous synchronization. This means, the synchronization for each segment received will be based on the half-way position and the frequency deviation ("phase velocity") of the received signals will also be estimated for each segment and corrected together with the estimation process and channel correction. Additionally, the receiver has to assign the received symbols, segments, etc., to the corresponding frame sections (control segments and data segments, certain of which numbers are given for each transmitter). This is done by analyzing the relative placement of the half-paths, and by finding and analyzing the control segments. Special algorithms are subsequently described that will be applied together with some alternative methods, in order to show that a system based on tolerances is feasible proposals. 2. 2 SYNCHRONIZATION OR TIMING -includes symbol synchronization, frame structure detection and some aspects of channel impulse response estimation. 2. 2.1 GENERAL APPROACH The synchronization of the symbols received and the assignment to the frame commonly used will be based on the correlation processes, that is, the corresponding sections of the received signal will be correlated with the reference or training sequence (half-way) stored in the receiver. For a general search or monitoring function, in order to detect all the components within the frame or grouping, this must be done on a symbol-by-symbol basis during at least one frame according to the following equation: v x + m Y m Q = x = N, m = 0 where x defines the position within a data sequence with Nfr = number of symbols per frame, vx + -. are the data values received, Ym represents the training sequence, and Cx distributes the amplitudes and phases of the signals received from segments. Figure 4 shows an example of a correlation result (only magnitudes), which comprises the components of three users (different magnitudes and constellations of delay). In order to assign the received component signals to the commonly used frame in order to detect the frame start, one or a combination of the following methods may be applied: Correlation results can be filtered over time in order to reduce the number of calculations: For example, up to 210 successive values will be removed and positions will be excluded as long as these values are at least 3 dB lower than the actual value; if a larger value appears, the last selected value and position will disappear if the value is at least 3 dB below the actual value, and the distance to the actual position is less than 210 steps; the counting process in this case is restarted. • The positioning relative to the magnitudes of the various segment signals derived by correlation are valued and correlated with the defined frame structure (different size and spacing of the control segments and general data or multiples of them, different magnitudes of the various users, etc.); however, this method, in some cases, may not give a clear result. • A section of successive data symbols, according to the relative placement of the identification number sequence of the cluster, or a fraction thereof, is selected together with each of the relevant correlation peaks, and a correlation is made between the chosen section and the corresponding reference sequence stored in the receiver; a suitable result, with respect to the result of the half-stroke, indicates a signal of control segments. • Finally, the content of all the signals received from the segment (or only from those that have been identified as being the control segment section) will be analyzed with suitable methods in order to detect the location numbers of the control segments, and so on.

It should be noted that very small components in the correlation result may come from others, for example, from distant networks, which use the same channel. These have to be or can be excluded from the additional evaluations by a more sophisticated processing, which takes into account that they may be outside the expected intervals of the means-journeys, vary in their relative position to the frame itself, no continuous reception, and they are not indicative via the frame control signals, "first segments" of the cluster itself. Other impulses to sporadic responses that include echo components, caused by an imitation (sub-optimal) of the half-path by the general data stream, may appear within the data sections. These impulses are also characterized by a variable position, from frame to frame, and additionally by the fact that there is always a largest main impulse within the same section. Figure 5 shows a correlation result where the components (same, in Figure 4) have been assigned to the used, detected frame. After a first analysis of the structure, that is, after knowing the start of the table and the placement of the segment, the scale x that is adequately displaced, the correlation process can be restricted to those intervals where the means are to be expected. -recurridos. To include concatenated synchronization deviations of the various signal sources, within a cluster, and caused by the different lengths and echoes of the route (see Table 1.4 in Chapter 1-4, "System Description"), the interval could be chosen to be between ± 9 μs or ± r = ± 75 symbols with respect to the previous frame in: Cl.sel - _, t-1-? - - r = l =. m = 0 xs are the selected central positions (note: xs differs by 8 from the corresponding position x in the first equation). The position of the value of maximum magnitude between | C?, Seil max = lmax, s within the range is called Imax, s- The reliability of the result with respect to the amplitudes and echo phases could be improved to a certain degree by performing a cyclic correlation, for example , of module 16, later, based on the maximum found in the previous process. The selection of the pertinent data section of a segment must be based on Ima?, S- The processing of the successive signal will be performed with the same local clock conditions in conjunction with the channel impulse response of the segment. The absolute deviation, remaining of the clock itself from an optimal (possible) position causes a deviation of the channel parameters, and is included in the correlation receiver Cx, se? and it will be automatically corrected in the corresponding channel correction process. A synchronization error, relative, remaining between the two devices, based on 25 * 10_d for each and calculated over the relevant interval of a segment, results in a synchronization error for the outer sections of the segments in the order of: ± 2 * 25 * 10 ~ 6 * 836/2 «± 0.021 symbols This can cause amplitude errors of less than 1.5% (yes function) and can be abandoned. 2. 2.2 OPTIONS, CHANNEL IMPULSE RESPONSE The processes defined so far can not give under certain conditions, a very clear maximum due to the following reasons: frequency deviations of the received and down-converted signal, causing reasonable phase deviations within the medium- travel, - sampling with greater time deviations (> 1/4-symbol) with respect to an optimal synchronization, causing a long end of lateral impulses (yes function).

However, the results should be good enough for an approximate synchronization, especially at frequencies (RF) of 2.4 GHz. On the other hand, an updated process might be necessary anyway, to estimate an approximate frequency deviation for a process start (see chapter 2.3, center frequency deviations and correction process) and the same method could be applied as well or the same results could be used for the synchronization and estimation of the phase impulse response. To overcome the disadvantage of • unfavorable symbol synchronization, a second correlation can be made within a time shift of half the sampling interval, which requires additional sampling with the inverted clock signal over at least the half-path plus one extension. Finally, the best constellation can be chosen to be validated for the entire segment (criterion: peak value or power remaining outside a certain time interval that is as low as possible). However, the effects sought with any of these measures are not too great, but without both, the correlation results could be degraded by more than 6 dB (compared to the optimal values) and a significant part of this can be recovered at apply them 2 . 3 OF SVIAC IONE S DE FRECUENC CENTRAL IA AND PROCEEDS OF CORRECTION 2. 3.1 APPROXIMATE ESTIMATION AND START PROCEDURE The central frequency deviations and the corresponding phase rotation values (per symbol as well as per data segment) have been investigated in chapter 1.2"System Description"); The relevant values for the reception are: Phase A deviation 2.4 GHz: «± 2.1 ° Per symbol A 5.7 GHz:« ± 4.2 ° These deviations over half the segment, the relevant distance between the middle of the middle course (correlation result) and the data symbols in the edge positions, are: At 2.4 GHz «(±) 870 ° A 5.7 GHz« (±) 1750 ° However, the deviation is a linear rotation in time of the complete signal that can be corrected by an opposite rotation of the received values, at least during the time of a segment.

The question is: can this be done in combination with a "closed" compensator, or is it a necessary combined compensation and correction process, and if so, how much complexity does this require? In the following sections, a correction method using compensator output corrector or intermediate values for calculating the phase / frequency correction values for the signal received at the input of the compensator is described. A simplified functional diagram of the process is shown in Figure 6. The baseband signal IN is fed to the input of a phase correction device 2, to the half-path correlation device 1 and a calculation and averaging device 3? F. The output signal of the phase correction device 2 is fed from an input of a compensator 5 (viterbi). The output of the compensator 5 is fed back via a signal reconstruction device 4 to the calculating device 3 and averaged? F. A correlation signal output 7 of the half-path correlation device 1 is fed to the control inputs of the compensator 5 and the signal reconstruction device 4. In the second output of the correlation device 1 Half-path is fed to a control input of device 3 of averaging f ?, as line 6 of the start condition signal. A starting position or value for the correction of the number of symbols around the half-way including the half-way itself, could be found by one or a combination of the following procedures: • Approximate estimation; Method 1: A series of optional correlations around the half-way is performed, either with the half-paths or the received data streams that are pre-distorted (in phase) according to certain frequency deviations (expected), and at the maximum it is determined in its position it produces a value for the "phase velocity"; • Approximate estimation; Method 2: The correlation process with the half-way (s) is divided into two halves; phase deviations between the main components of both sections are calculated on average; the result is valid for the distance of 8 symbols in cases of an individual training sequence of length of 16 or for 16 symbols in the case of two training sequences of length 16. Estimation of frequency deviation, improved; method 3: Alternatively or optionally, the sequence used in the control segment (s) for the identification of the cluster or a selected part of this sequence is used in conjunction with the channel impulse response and the results of either the method 1 or method 2, for the 2p module control, in order to calculate a more exact phase deviation value; the process can be divided into several calculations using the sub-groups of for example 8 symbols of the complete partition taken from the identification sequence of the grouping; or the results of the sub-groups are valid for the corresponding distance (s) between the section taken into account and the center of the training or mid-section section Although this method is correct only for the control segment (s) , the results, in the case of a limited Doppler influence, could also be taken as a basis for the other segment signals of the same user, in the same box Detailed description, method 1 Calculations during the correlation process have to be extended by a constant phase shift of ejfd per symbol, which results in a sequence, or displacement of jfd ie "d, where it defines the steps of the additional function , positive and negative values, and the position within the data stream (for example, positive and negative values with respect to the center of the half-way): C¡jtl -? vX? + l + mrmc1 «? -r = l = r This normally involves a complex conjugate multiplication of vx and Ym, but since Ym is only a sequence of real value, it returns to its normal multiplication. The correlation interval is defined by r, where similar conditions have to be chosen as in chapter 2.2 Synchronization, for example, ± 75 symbols. The results of the synchronization correlation can be used also or multiplied by ej f? . The process has to be performed during an appropriate number of d-values, frequency deviations. For example, at 2.4 GHz the deviation over the half-path is 34 °, and if this interval should be searched in steps of 10 °, equivalent to the resulting phase directions of 5 ° abs, or ± 2.5 ° with respect to the center, this will require 3 correlations of 2 times, positive and negative values. Of course, this amount can be reduced by a more sophisticated process. The maximum correlation result defines the deviation per symbol? = fd (with dopt) • Under reasonable noise conditions, the result could be improved by interpolating, for example, the three values around / inclusive to the maximum in order to determine the actual position of the maximum.

METHOD 2 There are going to be two different correlations with the training sequence divided into two halves, in the case of an individual sequence or with the two parts of a double sequence, for example 2 x 16 symbols, the results of the two Correlations are A + jB and C + jD. The relative phase difference (by symbol) between the two is: 1 (f AD - BL < Ps.r <? = Y = - arctan? «AC + BD where qx is given by the spaced amount of symbol between the centers of both sequences or parts. The phase difference between these sub-results should not exceed 180 °.

METHOD 3 The following steps are performed: The sequence used for the identification of the cluster, known by the receiver, is being modified according to a channel impulse response, convulsion process, in order to build a sequence identically to the section of the signal of each corresponding one, but without frequency deviation and the corresponding phase deviations; The direction and phase between the identification sequence of the modified group and the corresponding sequence of the received signal, is calculated symbol by symbol, by division or complex multiplication conjugated and averaged, the whole is identical to a correlation process, taking into account the possibility of two steps of module 2p included; finally the deviation per symbol is calculated. In order to estimate whether the module 2p steps are involved or not, the result of the estimate of the approximate frequency deviation, method 1 or 2, can be used to calculate a relevant, approximate value for the section investigated; defines the amount of the module steps 2p of integer that are included. The process of averaging usually has to be divided into several sub-processes with for example, sub-groups, of 8 symbols, in order to avoid estimation errors caused by phase deviations within the sequence taken into account.

The intermediate results of the subgroups are validated for the corresponding distances between the section taken into account and the center of the mid-course training section. These results are then divided by the corresponding distance in order to obtain a deviation by symbol. Finally, the various results are averaged. All methods suffer, more or less and although averaging, noise processes are involved (better conditions are achieved with method 3). However, the results in practical cases should be good enough to start a process as described later. 2. 3.2 FREQUENCY CORRELATION PROCESS In a first step, the data values received, starting from the first symbol (known) below and above the center of the half-way and continuing in both directions symbol by symbol, will be correlated with - (i-l / 2) ?, for the upper part and with - (- i + l / 2)? for the lower part (multiplications by ex (-? μi / 2)?. where i is the Index of the symbols starting with 1 for both sides of the center of the half-way and? is the result of the estimation of frequency deviation from above, preferably from the improved estimate). After this, the compensation process can be started, favorably also with the known, respective symbols of the half-way.

This produces, with a delay that depends on the length of the compensator or an appropriate part, of a number of corrected output values per anal for both sides from the center of the half-path. The compensator only knows the law of how the channel is constituted, but nothing about phase deviations as well as noise values, etc. However, the output values found by decision include a correction of the remaining small phase or deviations in frequency and noise, if the deviations are not large. In this way, if the output sequence and the channel law for reconstructing the input signal is used, these values will not contain the remaining phase deviations, and a division of both signals, the data values received vi and the reconstructed values yi, and the calculation of the corresponding phase angles must produce the remaining phase deviations? ±. : F?, = Arctan - E x-p are the output values of the compensator and hp represents the channel impulse response (identically to C?, se?) • In all cases, the corresponding values have to be used. For example, the last output value of the compensator of a sequence used to reconstruct the input value vx must also be marked with i. This implies that the reconstructed signal is delayed for a time equivalent to the depth of the compensated one. Assuming that this produces reasonable results, then can the values be used to update the values? by ±? i / i, from both directions and applying simple filtering. The new range is then used to pre-correct the next data values as before before entering the compensator, and so on. This applies to the entire segment. Another possibility is to maintain the value? from the approximate or improved estimate as? 0 and add only one additional component representing the additional correction according to? i, where a simple filter function can be applied: y = y0 + and con- > with rcorr ^ Q - ftrcorr-l + i where? COrr-? represents the previous correction value and Ycorr the actual value. ß defines the characteristic filter; a value of, for example, 0.05 could be used for a case in which the ß values represent groups or results of 8 averaged symbols. Due to the delay of the compensator, there is a kind of extrapolation included automatically, which can reduce the speed of the corrections and prevent the application of high Doppler effect frequencies, for example, in the mobile reception. The compensator decision errors and noise in general will probably have some influence on the correction process, especially at the beginning, but with the average error rates of up to 3 x 10 ~ 4, there should be no noticeable (additional) degradation. In any case, this or another similar strategy has to be confirmed by simulations of the complete system. 2. 3.3 ADDITIONAL OPTIONS Estimates of the frequency deviation elaborated for the various segments of one or several users could be averaged over a longer period of time and part of the The resulting deviation, for example, exceeds the value 1.5 times the tolerance allowed for a user could be used to correct the reference oscillator itself. 3. PROCESS OF ACQUISITION, MONITORING AND "PERCEPTION" OF THE CHANNEL 3.1 GENERAL MONITORING DURING THE SLEEP It is recommended, mandatory from the point of view of the network, that the devices that are not in use, depending on waiting, monitor at least the signals of the group itself or network, if present, in order to detect a message determined by itself, or, if a network does not exist, observe at least the other channels to detect a new start-up of a network of the same group and subsequently a message. This implies in the case of the detection of the network box and the evaluation of the control elements at least with respect to an announcement of new messages. In this case, the placement of the media-routes is approximately known, which simplifies the evaluation and reduces the required processing in a significant way compared to a situation where a network is not present and all possible channels need to be verified.

A current monitoring of the network itself, half-routes, allows the device to quickly acquire a corresponding part of the channel if requested. If a network of the group itself does not exist or only exists but a second one is allowed, then all the possible channels have to be identified with respect to the presence of signals, which implies the correlation with the half-way, determination of table, number of identification of the grouping and, if a network of the grouping itself appears, the verifications previously described have to be applied. The channels used by other groups could be excluded for a while but need to be checked from time to time in order to detect a release by the "old" user and an entry by a user of the same group. The monitoring normally includes the search for channels in the half-paths, which produces knowledge about empty channels and allows the device to be as fast as possible if it is asked to open a network. The processes are identical to those described in chapter 2.2. Synchronization / 2.2.1 during the general method. Whether it is a continuous correlation symbol by symbol or, in the case of a network of the group itself, a detailed analysis is carried out, where in these cases the magnitudes are relevant. A simultaneous process with different sequences (pre-distorted in phase) par.ece is not necessary. The corresponding formulas are: - for a monitoring process in general, applied to unknown channels: ft - v Y 0 = X = V. or continuously to monitor a channel used by the group itself or with a known frame, used by another group: where: Vx + .. are the data values received, Ym represents the training sequence, Cx and C?, sei are the correlation results, X (in the first equation) defines the position where at least the length of a table (Nfr) has to be investigated, Xs (in the second equation) are the selected central positions (note: xs differs by 8 from the corresponding position x in the first equation), and r defines the interval of correlation. Figure 7 shows an example of a correlation result obtained from a channel used by three transmitters, which may be a channel of the same group or another, a close grouping, depending on the identity number as described later, for which reason "*" marks empty segments; relevant only if the channel corresponds to the group itself. It is pointed out that some rather small components in the correlation result, among those that correspond to an identified group or network, may come from distant networks that use the same channel. This could be distinguished by a more sophisticated processing, taking into account that they may be outside the expected intervals of the half-courses and vary in their relative position to the known frame, no continuous reception. However, for monitoring and perception processes, these effects are of minor importance. It should be further noted that the correlation methods applied herein for the detection and evaluation of the signals adjusted to the system are inadequate to detect other signals such as interference from microwave ovens or amateur radio. Therefore, it is recommended to also estimate the power received in the relevant channels. In order to verify the identity of the grouping, the corresponding section (s) of the signal (s) of the control segment has to be demodulated, decoded and compared with the identity number of the own group. Alternatively, a correlation with a sequence according to the identity number can be made. Depending on the results and the result of the mid-course correlation, this produces different statements: • Positive verification identity, a half-way: signal (s) of the group itself, • Negative verification identity, but half-way: signal (s) of another group, • Negative result, no half-path for any one power, it can be interference sporadic In order to further verify the various signals received, the "first segment" indicated in the control segment could be evaluated. This provides an "indicator" for the corresponding data segments, positions that could be compared to the half-paths detected. However, the mid-course correlation must produce the most reliable information.

Claims (6)

CLAIMS:

1. A communication system that uses a TDMA time division multiple access data transfer method, where data is transferred in frames, with a table that is divided into a number of time segments and a table which consists of several control time segments at the beginning and the subsequent data time segments, where a specific control time segment is dedicated to a specific user and where in each time segment, a first time of security, defined, characterized in that, a frame alignment is made by the steps of: the synchronization of a first control time segment, busy of the frame is defined when referring to the synchronization of the last time segment, of control , occupied of the previous frame, each segment of time of control occupied, subsequent is placed in the time with respect to the previous segment of the time, of control, occupied, previous of the same square or, the synchronization of the time segments, of data, occupied within the table is defines when referring to the synchronization of the last segment of time, of the control, occupied, of the same frame, and where a defined second safety time is associated with the frame, this safety time that is placed at the end of the frame. wherein the second security time defined is extended with respect to the first security time, defined. The communication system according to claim 1, wherein for each signal sent in a time segment a half-path is added which is used to determine a synchronization reference point for the corresponding time segment. The communication system according to claim 1 or 2, wherein at a data time segment, a different security time is added -in terms of a time segment, of control, with any multiple of the security time of the control time segment that is different to the security time of the data time segment. The communication system according to claim 1 or 2, wherein two identical length training sequences are transmitted in the half-path of a time segment, in particular two identical sequences of length of 16 symbols. The communication system according to claim 4, wherein the two identical training sequences are directly joined together and provided to the outer edges with a number of cyclic information symbols and also with an "anti-symbol" number. ", continuing the cyclical information that has an inverted plurality. User station for a communication system according to one of claims 1 to 5, having a receiving unit and a transmitting unit, characterized in that, means are provided for evaluating the received signals that are sent in the time segments of control and data time segments and for determining the synchronization reference points for the control time segments, occupied and data time segments, where to transmit the control signals in the first control time segment of a frame synchronization, adjustment means are provided which adjust the synchronization of the first control time segment when referring to the synchronization of the last time, control, busy segment of the previous frame, where to transmit control signals in a control time segment, subsequent to frame synchronization is provided to the medium adjustment which adjusts the synchronization of the control time segment, subsequent to referring to the synchronization of the control, busy, previous time segment of the same frame, and where to transmit data signals in a data time segment of frame synchronization are provided which adjust the synchronization of the data time segment when referring to the synchronization of the last occupied control time segment of the same frame. SUMMARY OF THE INVENTION In a multi-user system for the wireless transmission of video, audio and / or combined signals or general communication signals, a proposal is given for the synchronization and timing of various messages of the users. The proposal is based on a combined transmission of FDMA / TDMA and an 'access technique. The data of the users are transmitted in a common frame comprising a number of control segments and also a number of data segments, each segment being associated with a security time. The segment signals have the characteristics that i) the position in time of the signal of the first control segment in the frame is based on the signal of the last control segment of the previous frame, ii) the position in time of the signals of the transmitted control segments is subsequently based on the signal position of the previous control segments, iii) the time position of the signals of the data segments is based on or refers to the signal of the last segment of the control in the same box, which can be a signal from another user or from itself, iv) at the end of the box behind the signal from the last segment a security time is placed additional or increased significantly greater than the general safety times associated with the other segments. The proposal also includes a communication system and a receiver for a communication system.