TWI334576B - Method and system for reading rfid tags - Google Patents

Description

1334576 P65950028TW 22800twf.doc/n IX. Description of the Invention: [Technical Field] The present invention relates to a method and system for reading a radio frequency identification tag (referred to as a still tag), and in particular (4) The method and system for the type-f stomach-type radio frequency identification (1 tag).

[Prior Art] Fig. 1 depicts a conventional RFID reader 1〇1 and three radio frequency identification tags 1〇2-1〇4. The reader 1〇1 sends an ultra-high frequency (ultra lion, UHF) or high frequency (HF) electromagnetic wave enabled tag 1〇2_104, so that the tags 102-104 can backscatter the tag signal to read Machine 1 (Π. The data sent by the RFID tag is modulated in this tag signal. In order to detect the data 'This RFID tag reader must receive this tag signal and then according to

This tag signal produces the above information. This data generation program is called signal detection, decoding or demodulation. By way of example, the radio frequency identification tags 102-104 can encode data using an FM0 modulation scheme. Figure 2 shows the data pattern of four FM0 modulation methods. In Fig. 2, there are four patterns S1 to S4. S1 or S4 represents the data symbol 1. S2 or S3 represents the data symbol 〇. According to the FM0 modulation mode, there must be a state transition between two consecutive symbols (from low state to high state or high state to low state). Figure 3 is an example of a standard signal waveform according to the FM0 modulation method. Waveform 301 is a completely noise-free label signal waveform. 5 1334576 P65950028TW 22800twf.doc/n There are two traditional methods for signal detection, namely edge detection and matched filtering. The edge detection detects the state transition of the tag signal waveform through the intermediate line 305, and then reconstructs the data bit according to the state transition statistics. For example, the data symbol 丨 has no state transitions during its symbol period, while the data symbol 0 has a state transition in the middle of its symbol period. Edge detection is only useful when the tag signal waveform has relatively low noise. As shown in Figure 3, waveform 302 is a tag signal waveform with noise from a distance of 2 meters and carries the same information as waveform 30丨. Although waveform 302 is distorted by noise, its data pattern can still be reconstructed by edge detection. County - an example to illustrate, the waveform 3 〇 3 is the other ship 8 meters away from the distance, ▼ the noise of the tag signal waveform, still carry the same information as the waveform. However, severe noise causes waveform 303 to have many narrative transitions across intermediate line 305. Therefore, the edge detection will misinterpret the data pattern. Matching filtering has greater noise tolerance than edge detection. Match overfishing = in the data _ SM4 and the tag signal waveform __ match ^ yuan. Fine, in order to see the correct cylinder, the pulse rate of the tag signal must be known, and must be happy. The data of the offerings: the situation is easy to obtain, especially the printing type ___ printing type of aging is actually (four) ink printing system is cheaper than the implanted standard iron. However, 歹 = up to 2: has a non-f-stable data clock rate. The amplitude drift is very typical. Moreover, the data rate of the data when you move a symbol is 6 1334576 P65950028TW 22800twf.d〇c/n. The clock rate may be different from the data of the next payment. Because the label signal of the frequency identification tag carries noise, and the edge detection is used _^2 = [Summary of the Invention] The present invention provides a method for taking the radio frequency identification tag to be more effective. 6 shells

The present invention provides a problem of clock recovery, synchronization, and data frame decoding of a noise-receiving tag from a printed RFID tag. Therefore, traditional technology is more effective. Death ratio

Filtration techniques are not feasible. The present invention provides a method of reading radio frequency identification tags. This method includes the ^ step. Therefore, receiving a tag signal from the difficult sign. The data clock rate is returned according to the pulse length of the tag signal. Then, according to the data clock rate, a signal correlation calculation (signal correlati〇n) between a preset signal pattern (sig(10)1 patter) and a preambie in the label signal is performed to decide to follow the previous guide in the label signal. The synchronization point of a data frame of the data. Finally, using the adaptive Viterbi algorithm to decode the data frame on the extended trellis diagram, the nodes and branches on the extended trellis diagram are based on the data clock. Possible changes in speed. The invention further provides a system for reading a radio frequency identification tag. The system includes a receiver, a reply unit, a synchronization unit, and a decoding unit. The receiver is connected to 7 P65950028TW 22800twf.doc/n. The reply unit handle is connected to the receiver, and the second and the second v are signed, and the pulse length is equal to the data of the thief signal. The synchronization unit is lightly connected to the reply unit for predicting the clock speed according to this data. The 'U-cut operation between the signal pattern and the leading data of the standard signal to determine the time at which the previous information is followed in the label signal. The code unit is used to expand the map. The adaptive Viterbi algorithm is used to decode the data frame, and the arrangement of the point and the branch of the expanded sub-picture is based on the possible change of the clock rate of the data. [Embodiment] Times, in the typical label signal The RFID tag repeats the transfer, and the tag data of each repetition is included in the data frame. The first two materials guide each data frame, and each of the preamble data and data frames contains two records. _4 is the reduction (four) - the actual _ private reading and shooting = the focus of the method flow chart H receives the label information from the radio frequency identification: step 410) 'and then reply to the label signal data rate (ie two, 420 ). After the data clock rate is identified, the synchronization step is performed, = the position of the frame = point between the preamble (4) and the data frame following the material data (ie, step 43G). The synchronization point of this frame is the leading data: the point where the east and the data frame start. After identifying the data clock rate and frame point, the process will continue to decode the data frame and obtain the data (ie, step 440). Although the circuit difference causes the data of the printed RFID tag 1334576 P65950028TW 22800twf.doc/n

Rate drift, (4) This drift depends on _ slower, and the method of this embodiment can be adapted to the clock. Therefore, the resource clock rate replied by Lai Xing can be used to calculate the signal step. The data rate of the reply data can also be used as the most inferior frame after the synchronization of the clock scale. The steps of Figure 4 are described in detail below. In the present embodiment, the clock recovery step (step 42 of Fig. 4) can be further broken down into four steps 421 to 424 of Fig. 5. First, a signal histogram is generated based on the pulse length of the standard signal (ie, step 421). Take the example of FM0 modulation as an example. Here - a pulse represents the part of the tag signal that has no state transition (ie, does not cross the middle line). The pulse length represents the length between two consecutive state transitions of the tag signal. Figure 6A shows a signal histogram of a tag signal with noise from a fingerprint RFID tag. The signal rectangle of Figure 6A shows the probability statistics of the pulse length of the tag signal. Signal The vertical axis of the histogram represents the probability of a probability mass fimction (PMF) from the pulse length. The horizontal axis of the signal rectangle represents the pulse length. The unit of the pulse length is the number of sampling points within one pulse. The RFID tag reader samples the tag signal at a predetermined fixed interval. As shown in Figure 6A, the most common occurrence in the signal histogram is a very short pulse of one sample point due to severe noise. After generating the signal histogram, the flow proceeds to generate a plurality of pattern histograms (i.e., step 422). Each of the graphic rectangles is generated according to the pulse length statistics of the modulation signal (in this embodiment, FM0) conforming to a preset clock rate, and these graphics are 1334576 P65950028TW 22800twf.doc/n rectangular diagram The preset clock rate is selected within the range of the preset rate interval. For example, three kinds of graphic rectangles of the prosthesis under the methods of FIGS. 6B to 6D are illustrated. Each of the graphs of Figures 6B to 6D == swells to the respective preset clock rate. Such as _ 2 with

There are two kinds of pulse patterns in FMO modulation, that is, the probability of short pulse and long pulse ^ bean pulse is long pulse _ twice. If the corresponding one-shaped rectangular preset clock rate is the shooting _ take the machine to turn Four times, = the resulting graphical rectangle is shown in Figure 6B. If the preset clock rate corresponding to a 1-shaped rectangle is six times the sampling frequency of the RFID reader, the resulting rectangular graph is shown in Figure 6C. If the preset clock rate corresponding to a graphic moment y map is eight times the sampling frequency of the RFID tag reader, the resulting graphic rectangle is as shown in the ® 6D. It is recommended to pre-set the above preset clock rate. Set the range to the range of possible variations in the clock rate of the data containing the printed RFID tag.

After generating the graph rectangle, the process continues to compare the signal rectangle and each graph rectangle (ie, step 423), and then selects the preset clock rate of the graph rectangle closest to the signal rectangle as the data clock of the label signal. Rate (ie, step 424). For example, in the present embodiment, a graphic histogram having a minimum matching error in all the graphic histograms can be selected. For better clock recovery' In other embodiments of the invention, the no-noise pulse on the graph rectangle can be replaced by a Gaussian profile based on a no-noise pulse. The graphic rectangle with the Coats profile is more convenient. 1365576 P65950028TW 22800twf.doc/n It is easy to match the signal rectangle with noise, which can restore the data clock rate more accurately. Further, in other embodiments of the present invention, since the graphical histogram is not associated with an individual target signal, the graphical histogram can be generated prior to receiving the individual tag signal and reused after the manuscript. In this case, after the graph histogram is generated, step 422 can be ignored in the next tag reader. Fig. 7 is a comparison of two clock recovery methods and two conventional methods according to the present embodiment. The vertical axis of Figure 7 is the mean square error (MSE) of the symbol period replied by each method. A lower mean square error represents better clock recovery performance. The horizontal axis of Figure 7 is the signal-to-noise ratio (SNR) of the received tag signal. The lower signal-to-noise ratio represents a more noisy tag signal. Figure 7 compares four methods, the first trial (FirstTrial), the majority (Majority), and the simple statistical method (Si_e).

Statistics) and Advanced Statistics. The first test method is a conventional method in which the pulse length of the first state transition of the tag signal is taken, and the reciprocal of the symbol period is calculated as the data clock rate. This method is very simple and can only be operated with a stable and high signal-to-noise ratio. However, for printed RFID tags, it is almost impossible to have stable and high signals. ^ Most methods are another traditional method. The method is to obtain multiple pulse lengths between the transitions of the two-person state of the tag signal, and then calculate the symbol period by the majority vote (rzo„nty vote) or the average of the pulse lengths. Then count the reciprocal of the sign period and It is specified as the data clock rate. This method is more complicated 'When the wire is in the middle and upper level performance _. However, if the waveform is not 3G3', when the signal-to-noise ratio is very high, the serious noise leads to 11 11334576 P65950028TW 22800twf.doc/n Pseudo-transition of the multi-number state 1 - Both the test method and the majority method are poor due to the pseudo-sense, resulting in a short pulse. The simple statistical method according to the embodiment of the present invention contains none. The clock recovery method of the graph rectangle of the noise pulse, and the advanced statistical method according to another embodiment of the present invention is a clock recovery method including the graph rectangle of the Coats profile. As shown in FIG. root The method of the embodiment can be superior to the traditional method in the case of low or high signal-to-noise ratio. In addition, the performance of the advanced statistical method is better than the simple statistical method. After the data clock is replied, the next The stage is to calculate the frame synchronization point of the data frame following the preamble data in the tag signal. In this embodiment, the signal correlation operation is performed between the preamble data and a predetermined sigign pattern calculated according to the data clock rate. More specifically, the preset signal pattern is a preamble pattern defined by a standard specification followed by the printed RFID tag. In addition, the signal correlation operation used in the embodiment is a window sliding correlation operation. (wind〇w-sliding correlation). Since the frame synchronization point is the end of the preamble data and the beginning of the data frame, this embodiment uses the signal correlation operation to determine the position of the preamble data in the tag signal, and then can determine the tag signal to follow. The location of the data frame of the leading data. Referring now to Figure 8, Figure 8 shows the leading profile graphic 800 defined by the standard specifications. And preamble data waveforms 801-803 received from the tag, waveforms 801-803 are based on three different data clock rates. The preamble data waveform 800 is defined in standard specifications and is well known to those of ordinary skill in the art. The leading data 801~803 and the leading data graphic 800 phase 12 1334576 P65950028TW 22800twf.doc/n • Similar. = The difference between the information 801~803 and the leading data graphic 800 is the miscellaneous information and the information Wei drift material. The window sliding correlation operation ^ refers to the leading data waveform received by f from left to right and across the standard (4) data pattern 8GG. The best match between the two sides of the data bribe (four) begins. The data clock rate replied at step 420 is only an estimate of the data clock rate of the preamble data received by the reader. Even if there is some mismatch in the data rate of the data of the preamble data actually estimated by f, the φ-sliding correlation operation is still valid. However, if the above estimation differs too much from the data rate of the actual received preamble data, the window sliding correlation operation will fail. This is why the data clock rate must be estimated before the data frame is synchronized. The clock rate of the step gamma estimate provides a good reference point for signal correlation operations. Please note that the length of the standard lead data graph 800 may have a decisive influence on the accuracy of the synchronization. The longer the length of the leading data pattern _, the higher the accuracy of the data frame synchronization. Fig. 9 is not a conventional synchronization method and a comparison according to one of the embodiments (5). The vertical axis of Fig. 9 is the square error of the synchronization offset of the linear scale. The horizontal axis of Figure 9 is the signal-to-noise ratio of the received signal. The power threshold method (p〇wer ^hreshold) ▲ is a traditional synchronization method, which is to treat the human character '4 transition of the received label signal as the starting point of the data frame. In the comparison of Figure 9, there are two variations in the power critical method. The first variation uses a four-symbol length to synchronize the tag signal. The second variation uses a preamble of length ten symbols to synchronize the tag signals. A correlation algorithm (Correlator) is a synchronization method according to the present embodiment. Similarly, in the comparison of Fig. 9 13 1334576 P65950028TW 22800twf.doc/n, there are two variations of the correlation algorithm. The first variation uses a leading data of four symbols to synchronize the labeled signals. The second variation ^ synchronizes the tag signal with a preamble of length ten symbols. As shown in Figure 9, the performance of the correlation algorithm is much better than the power critical method, and the preamble data performs better. x W tongue ^ after the track frame synchronization point 'down' - the step is to reconstruct the data from the tag for decoding the data frame j. In order to overcome the noise and data clock drift, the third embodiment uses an adaptive Viterbi algorithm on the extended grid map. The extended trellis diagram is an extension of the traditional trellis diagram. For example, FIG. 1A: does not have a conventional grid map 1000, and FIG. 11 illustrates a sub-picture graph that is performed by a conventional FMO grid map. The extended plaid diagram includes a plurality of sections 1102) a branch line of two or more connected nodes (for example, a branch line), and an arrangement of the center point and the branch line is according to a modulation scheme of the data frame (FM0 in this embodiment) and Possible changes in the data rate of the data. As shown in Fig. 11, the nodes in the extended trellis diagram form a plurality of super ^ = super nodes containing three nodes, such as super = sub-picture 1 just super node is similar to the traditional grid surface = : The arrangement is based on the data frame modulation method possible on the other hand: on the other hand, each super node has a possible change above the node row. Among the super nodes, the clock rate of the material clock rate minus the head clock rate of the head line is the default clock rate unit of the symbol. For example, when the preset (4) + the node in the middle represents the current symbol data, 14 1334576 P65950028TW 22800twf.doc/n The previous symbol is the same. The node below the super node represents the data clock rate of the previous symbol plus the data rate of the previous symbol plus the preset clock rate unit. In this case, the Vitalbi algorithm can solve the problem of noise and data shifting. ,

The super node on the extended trellis diagram 1100 can be extended to include multiple nodes. For example, each super node may include five nodes ·', = 3 11 7 three nodes. In a super node with five nodes, the node of the field represents the current symbol of the dragon. The time of the ride is 1% of the pulse rate minus the two preset clock rate units, the identification, ', the current symbol of the table The data clock rate is the previous symbol pulse = a preset clock rate unit, and the other three nodes are the same. And so on, each super node can ^ more f nodes to accommodate a larger data clock drift range. There must be a one-to-one correspondence between pulse rates. Moreover, the foregoing must be within a predetermined range according to the preset rate of capture = the second clock rate must not be covered by the data clock during the period of the data frame. Adapt to the timing changes of the frame synchronization point. The decoding rate ^^ clock rate drift and ^ one step is (four) - a super" point when a step is mixed. In the expansion of the grid axis m (four) (four); ^ point = 15 1334576 P65950028TW 2280〇twf.doc / n (trellis _) In the middle of the road to find the smallest error is not the road control. The second step is in the first grid = thick line mark connected to all the grid road ", find: the most: the last marked path _. The side symbol is the leftmost indicated by her child's amount

The η of the method, according to the decoding method of the embodiment and the two traditional parties: the horizontal axis of the =12 is the matching of the received standard iron signal, ',,'; the consideration is the previous prior art part The conventional decoding method mentioned. The adaptive Viterbi algorithm is the decoding method according to the embodiment. As shown in Fig. 12, the adaptive Viterbi algorithm has the lowest whether the signal-to-noise ratio is low or high. The bit error rate. In addition to the method of reading the radio frequency identification tag in the above embodiment, the present month also includes a system for receiving the radio frequency identification tag. Referring to FIG. a, the system Goo is another embodiment according to the present invention. The system includes a receiver 1301, a reply unit 13〇2, a synchronization unit 13〇3, and a code unit 1304. The reply unit 1302 is coupled to the receiver 13〇1. The synchronization unit 1303 is coupled to The reply unit 13〇2 is coupled to the synchronization unit 1303. The receiver 13〇1 receives the label information from the radio frequency identification tag, and the reply unit 1302 statistically responds to the label signal according to the pulse length of the label signal. Clock rate. The synchronization unit 1303 performs a signal between the preset signal pattern and a preamble data in the label signal according to the data clock rate. The resolution of the signal is 16 334 576 368 P f f 疋 疋 疋 疋 疋 疋 疋 疋 疋 ^ ^ 跟随 跟随The data ^1 box should be dimensioned = the ratio of the method is decoded by the Wei box. ^ In fact, the system 1300 follows the method of labeling the process as shown in Figure 4. Figure m ^ needs to know m 〇uw 瓜瓜矛王- There is a close corresponding md '°n 301 between the systems 1300 to perform step 410. The reply unit 1302 executes the step execution synchronization unit 1303 to perform step 430. The decoding unit 1304 executes the (10)\ because the details of the steps 410 to 440 have been simplified in the previous embodiment' and will not be described again. According to the comprehensive fall, the method and system provided by the present invention use the data rate of the statistical statistics of the Weidai signal, using the signal pattern correlation 2, the synchronization point position 'and the adaptive Viterbi continuation method in the extended lattice map. Information frame. Therefore, the method and system provided by the present invention have the effect of decoding the noise signal from the printed RFID tag. The adaptive Viterbi algorithm can to some extent compensate for the error of the data clock drift and the synchronization point of the frame. Therefore, it is not necessary to calculate the time scale of the data in the recovery phase and calculate the time of the bet in the time phase. This helps overcome the problem of drifting the data of the printed RFID tag: drift. The performance of the method and system provided by the present invention approximates the maxjmum dec〇der, ml decoder. The method and system provided by the present invention are much less complex than the maximum likelihood decoder, since the method and system of the present invention only tracks symbol boundaries, rather than the calculations performed by the maximum likelihood decoder. The boundary of a symbol. In addition, although the method and system of the above embodiment are only used in the reading and printing of the prints, the present invention can be read on the 雠 义 _ , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , , For example, ^ is not printed _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ It is not intended to limit the invention, and any person skilled in the art can make a few changes without departing from the spirit and scope of the invention. Therefore, the scope of protection of the present invention is attached to the scope of the patent application. BRIEF DESCRIPTION OF THE DRAWINGS [Fig. 1 is a schematic diagram of a tag reader and a radio frequency identification tag of a conventional RFID tag system. Fig. 2 is a data symbol waveform diagram of an FM0 modulation mode. A noise-free waveform and two noise-carrying waveforms of a typical signal from a radio frequency identification tag. Figure 4 is a flow chart of a method for reading a radio frequency identification tag according to an embodiment of the present invention. Figure 5 is an embodiment of the present invention. Figure 6A is a signal rectangle diagram of a received signal with noise in an embodiment of the present invention. Figures 6B-6D are graphic rectangles of a received signal without noise in an embodiment of the present invention. 1334576 P65950028TW 22800twf.d〇c/i transmission == clock recovery method and two data patterns according to an embodiment of the present invention and FIG. 8 illustrates a standard preamble of received preamble data according to an embodiment of the present invention. Waveforms Figure 9 illustrates a comparison of a method in accordance with the present invention.

Figure 10 Figure 11 Exhibition grid. A grid diagram of a conventional FM0 modulation method is shown. The method for expanding the FM modulation mode according to an embodiment of the present invention is as follows: L illustrates a decoding method according to an embodiment of the present invention, and a system for reading the identification of the secrets of the traditional system. Main component symbol description] • 101: RFID tag reader 1〇2_104: Radio frequency identification tag S1~S4 · Barrier symbol pattern 301~303: Tag signal waveform 305: The middle line of the tag signal is 410~ paste, 421~ 424: Flow chart for reading the screening method Step 800: Standard leading data pattern 801 to 803: Waveform of the received leading data 1334576 P65950028TW 22800twf.doc/n 1000: Grid diagram 1010 of the conventional FMO modulation method: Conventional trellis path 1100: an extended trellis diagram 1110 of an FM0 modulation method according to an embodiment of the present invention: a trellis path 1101 of an extended trellis diagram: a node 1102 of an extended trellis diagram: a branch 1103 of an extended trellis diagram: a super node 1300 of an extended trellis diagram : System 1301 for reading a radio frequency identification tag: receiver 1302: reply unit 1303: synchronization unit 1304: decoding unit

20

Claims (1)

P65950028TW 22800twf.doc/n X. Patent Application Range: 1. A method for reading a radio frequency identification tag, comprising: receiving a tag signal from the radio frequency identification tag; according to a pulse length of the tag signal, a complex data rate In addition to the data, the clock rate is 1 and the signal is calculated between the data to determine the data frame in the target oxygen = = follow the leading data Position of the synchronization point of the frame—δ; using an adaptive Viterbi algorithm on the extended trellis diagram to solve the arrangement of nodes and branches on the extended trellis diagram. . '方」:! Please please? (4) The step of reading the RFID tag of the radio frequency identification tag described in item 1 includes: the pulse length of the signing signal generates a rectangular chart of the signal; (4) the graphic rectangle of each of the vines The average of these graphic rectangles should be: the pulse length of the modulation signal of the preset clock rate is 1°, and the speed of the pulse is in the preset range. Selecting the interval; = comparing the signal rectangle to each of the graphic rectangles; and k selecting the preset clock corresponding to the graphic rectangle corresponding to the signal rectangle to be the scale rate. (7) The scope of the read-only identification tag in the patent range 帛2 item is the range of the clock rate of the RFID tag 21 1334576 P65950028TW 22800twf.doc/n. 4. The method for reading a radio frequency identification tag according to claim 2, wherein each of the graphic rectangles represents a pulse length corresponding to a modulation mode of the signature pulse rate of the preset clock capture rate. No noise probability statistics. 5. The method of reading a radio frequency identification tag as described in claim 2, wherein each of the graphical rectangles includes a corresponding clock rate The Coats profile generated by the no-noise probability of the pulse length of the modulation signal of the tag signal. 6. The method of reading a radio frequency identification tag according to claim 1, wherein the preset signal pattern is a preamble data pattern defined by a standard specification followed by the radio frequency identification tag. ^ 7. As in the case of applying for a patent (4), the method of reading the singularity of the singularity is the window sliding correlation operation. 8. The method of reading a radio frequency identification tag as described in claim </ RTI> wherein the plurality of nodes constitute a plurality of super nodes, and each of the plurality of nodes comprises a plurality of the nodes. ° 9. As described in the scope of application for patent application 帛8, the frequency of the reading and decoding identification of the standard ^', the arrangement of the super nodes is based on the modulation of the data frame 10. As claimed in the eighth paragraph of the patent scope Each of the super nodes in which the radio frequency identification tag is read is arranged as a possible change in the root clock rate. Bay 11. As stated in Item 8 of the application, the readings of the superscripts, the nodes of each of the super nodes and the plurality of preset clocks 22 1334576 P65950028TW 2280〇t\yf.doc There is a one-to-one correspondence between the /n rates. Method H~Specially she encircles 11 items to read the shed identification mark between the two rates: set the clock rate to be within a preset range according to the pre-determined method range Cut label clock rate: can:: range contains the data during the data frame period Fang,,#·,4=u The steps of decoding the data frame of the 8 readings and readings of the index are further included: (4) Compared with I, the super nodes are regarded as a single node. And in the expansion - each of the super nodes on the first trellis path is taken - that is, the traitorous smear that is connected by the point, correcting the most sub-path; and the younger brother - the second trellis path of the frame 2 A symbol of the indication is used to decode the data. The system for reading a radio frequency identification tag comprises: a receiver for receiving a tag signal from a radio frequency identification tag. - a reply unit, secret to the receiver, for The statistical value of the pulse length of the label replies to the data rate of the label signal; a sync unit coupled to the reply unit for performing an initial signal pattern and the label signal according to the capture rate Between the hope and the other, 23 1334576 P65950028TW 2280〇twf.d〇c/Q is calculated to determine the position of the frame synchronization point of the frame with the _ preamble in the tag signal; and your field tf70 And coupled to the synchronization unit, using an adaptive Viterbi algorithm to solve the data frame on an extended trellis diagram, wherein the node and the branch line (4) on the extended trellis diagram are based on the (four) material clock speed Variety. For example, the system for reading a radio frequency identification tag as described in claim 15 of the patent scope&apos;, the reply unit generates a signal rectangle according to the pulse length of the tag. 17. The system for reading radio frequency identification of oxygen as described in claim 16 wherein the reply single god generates a plurality of rectangular shapes, each of which is based on a corresponding-predetermined clock rate. The pulse length of the modulation mode of the tag signal is generated by counting the pulse length. 18. The system for reading a radio frequency identification tag as described in claim 17 and wherein the predetermined clock rates are selected within a predetermined range according to a predetermined rate interval. 19. The system for reading a radio frequency identification tag according to claim 18, wherein the predetermined range is a possible variation range of a data clock rate of the radio frequency identification tag. 20. The system for reading a radio frequency identification tag according to claim 17, wherein the response unit compares the signal rectangle to each of the graphic rectangles and selects the graphic closest to the signal rectangle. The preset clock rate corresponding to the histogram is taken as the data clock rate of the data. 21. The system for reading the radio frequency identification tag 24 1334576 P65950028 TW 22800 twf.doc/n as described in claim 17 of the patent scope, wherein the graphic rectangles represent the manner of the tag signal corresponding to the preset pulse rate. Pulse length no-noise probability statistics. / 22· Read the radio frequency identification tag as described in the scope of claim 4, wherein each of the graphic rectangles includes a pulse length corresponding to the modulation mode of the tag signal of the preset clock speed ^ The Coats profile generated by the statistics of the probability of the wire. 23. The reading of the radio frequency identification standard as described in Item 15 of the Japanese Patent (4), wherein the preset signal pattern is a leading data pattern defined by the standard specification of the RFID tag. Xi Xi Shen 4 specializes in her reading of the 23-part reading and lying label '1, where the signal-related operation is a window sliding correlation operation. The preparation of the frequency conversion identification label as described in item (4), 糸, wash, wherein the nodes constitute a plurality of super nodes, the super node finds the home health, point. 2 = 6. As for the reading of the 25th item of the full-time enclosure, the array of the super-screams is arranged according to the tune of the data frame. - The reading of radio frequency identification tag data as described in the scope of patent ^25, the nodes of each of the super nodes are arranged according to the possible variation of the rate of the poor 4 pulse. The wire 8 is read by the radio frequency identification tag as described in Item 25 of the application of the fourth aspect, wherein each of the nodes of the super nodes has a corresponding correspondence with a plurality of preset time rates. relationship. For example, the system for reading the axis identification label 25 1334576 P65950028TW 22800twfdoc/n, wherein the preset clock rates are selected within a preset range according to the preset rate interval. The system for reading a radio frequency identification tag as described in claim 29, wherein the predetermined range includes a possible change in the clock rate of the data during the data frame period. Μ For example, the reading of the shooting label mentioned in the 25th item of the special observation section of May's towel will be regarded as a single item for each of the track nodes and the corresponding one in the extended lattice ® _ ~ The fourth (4), the wrong H (four) path of the state transition, and then the mother of the road k - each of the super nodes takes one from 4 horses of the information t all the lattice path towel, find the smallest error H connected The second lattice road is given a symbol as a solution. 26