JP4180641B2 - Method and device for data communication in a multi-user system - Google Patents

Description

  The present invention generally relates to methods and devices for performing data communications in a multi-user system.

  For many applications, a fast, efficient and reliable means of performing data communication in a multi-user system is desirable. Such a method is required when multiple data (from multiple sources) need to be quickly read by the receiver. One specific application of such a technique is used in electronic identification of multiple items.

  The electronic identification industry is an important industry for many commercial and military applications, including real-time item tracking and inventory management. By utilizing the electronic identification function, operational efficiencies can be greatly increased in a myriad of schedule plans, including virtually all schedule plans for certain forms of manufacturing, warehousing, distribution, and retail. If accurate real-time inventory tracking can be performed quickly and efficiently, various forms of waste, such as misplacement of products, overstocking or shortage of products, and product theft can be greatly reduced.

  Currently, the electronic identification industry relies heavily on manual (using light) scanning to assign a product code to each product to identify multiple products. Currently, universal product code (UPC) systems are widely used throughout the US retail industry. However, manual scanning of merchandise is very time consuming and prone to human error.

Therefore, an efficient and highly reliable method for transmitting data from a plurality of transmission sources to a receiver is required.
US patent application Ser. No. 09/981031 US Patent Application No. 09/978890 US patent application Ser. No. 09 / 982,271 US patent application Ser. No. 09 / 98,279 US patent application Ser. No. 09/981476

  The present invention discloses an improved communication method for quickly and efficiently communicating information from multiple source devices to a destination device.

  The communication system to be described achieves performance superior to that of the prior art by using a plurality of methods together. In accordance with the present invention, UPC replacement means can be utilized while adding further features and advantages, such as eliminating manual (using light) scanning and significantly increasing the scanning (or product identification) speed. Furthermore, in the present invention, it is possible to simultaneously identify a large number of products, and there are numerous applications such as inventory and retail inspection.

The present invention will now be described by way of example only with reference to the accompanying drawings in which like numerals indicate like elements.
In preferred embodiments of the present invention, the source device generally uses a one-way communication (from the source device to the destination device) to simplify the circuit of the source device, so that the source device uses a receiver. There is no need.

  Information communicated from the source device to the destination device typically takes the form of a binary electronic product code (“EPC”) or identification (“ID”) information, but in any way these It is not limited to the information format. It is also possible to communicate other types of information such as electronic telemetry (or other types of measurements or assigned data). Indeed, in the present invention, it is possible to convey information having a representation in binary (or other) number format.

  As shown in FIG. 1, information is typically communicated from a set of source devices 110, 120, 130 to a single destination device 100, and in the preferred embodiment of the present invention, a set of sources. Simultaneous information communication from the device 110, 120, 130 to the destination device 100 is used. Since the present invention has various applications depending on the contents of the embodiments, the terms used throughout the present description may be replaced with other terms for the sake of clarity. Thus, terms such as source device, transponder, user, product, tag, etc. can be used interchangeably throughout the following description without loss of generality, and can also be used as destination device, system controller, query Note also that the terms interrogator, reader, receiver, etc. may be used interchangeably throughout the following description without loss of generality.

  The communication system employed in the present invention can include several types of communication forms 140 such as optical communication, radio frequency (RF) communication, wired (contact) communication, electrostatic coupling communication, or inductive coupling communication. Although the preferred embodiment of the present invention utilizes a capacitively coupled communication link between the tags 110, 120, 130 and the reader 100, other forms of communication links can be utilized without limitation. It is.

  The following detailed description of the present invention is divided into five main sections that illustrate important aspects of the system and a final section that provides examples of system operation. In the preferred embodiment of the present invention, all the important techniques described below are used, but in other embodiments, only a subset of the techniques described can be used.

I. Data Scramble and Scramble Analysis As shown in FIG. 2, the data 200 communicated to the reader 100 by the tag 110 in the system being described is measured data or other user-defined data as described below. It can take various forms such as data. In a preferred embodiment of the invention, the transmitted data 200 consists of at least one identification data sequence. For example, the data 200 is composed of at least one EPC having 96-bit identification data. An outline is described in “The Electronic Product Code” (MIT-Auto ID Center, January 2001) by David L. Block. The EPC 200 is used to uniquely identify each tag (or product) 110 in the system by means of a header 203, an object class 204, a vendor code 205, and a serial number 206. For example, 96-bit information gives a huge number of unique IDs (from 2 ⁹⁶ to 8 × 10 ²⁸ , how huge this number can be seen from the fact that the Earth's mass is 6 × 10 ²⁷ grams). Note that it is obtained.

Typically, supplemental information 202 such as user information, error checking or correction information (eg, forward error signal correction (FEC), cyclic redundancy check (CRC), etc.), and other reserved bits is on the tag 110 of the preferred embodiment. Is added to the data 200 stored in. Supplementary information (eg, error detection or correction data) can be added before and after the data scrambling process described below, but if this supplemental information is added after data scrambling, a more uniform random characteristic. Please note that

  Those skilled in the art will also define a plurality of different information forms (eg, programmable time stamps, other user personal identification numbers (PINs), measurement data, environmental data, etc.) in advance, and tags 110, 120, 130. You will understand that it can be stored in Note that no limit is imposed on the amount or type of data stored in the tags 110, 120, 130 of the described system.

  All tag functions are typically implemented with low complexity (ie low cost) circuitry. In order to simplify the circuitry on the tag 110 and to enhance the performance of the channel selection process in the system (detailed below), it is highly scrambled to scramble the original ID data 200 before storing it in the tag 110. desirable. This is typically done by a randomizing or scrambling step 211 that is performed prior to the operation of storing the data 230 in the tag 110.

  This scramble algorithm 211 is typically applied universally throughout the system, ensuring that the EPC data 200 exhibits desirable statistical (ie, uniform and random) characteristics after being scrambled (220). The Alternatively, in other embodiments, some other scrambling, encryption, or numbering algorithm can be applied to stored data 200 to effectively generate scrambled data 220. In order to protect the privacy of the supplemental information, individual manufacturers may optionally apply the pre-encryption process 210.

  FIG. 3 illustrates an example system for embedding scrambled data 220 in a tag 110 according to a preferred embodiment of the present invention. In FIG. 3, the original EPC 200 is obtained in a normal manner from an EPC manager 310, such as a manufacturer. Thereafter, the EPC 200 is input to a scrambler 330, a scramble algorithm is executed by the scrambler, and scrambled data (S_EPC) 220 is output. Next, the scrambled data S_EPC 220 is embedded in the tag 110 by the RF tag programmer / writer 350. The scrambled data 220 is a modified version of the original data 200 and is now placed in the tag 110.

  FIG. 4 shows a high-level block diagram for simultaneously reading the electronic identification data 200 from a number of RF tag devices 110, 120, 130. This example describes a method for reading a tag associated with a product placed on a shelf during normal inventory. When activated, the reader 100 simultaneously activates a set of tags 110, 120, and 130 simultaneously. The activated tags 110, 120, 130 then continue the multipath transmission algorithm using the scrambled tag data 220 as the basis for channel selection (detailed in Section III below).

For example, in the first path of the multi-path algorithm, at least part of S_EPC1 (embedded in tag 110) is used to select channel A 240 and at least part of S_EPC2 is used to select channel B 240 and select channel C 240 using at least a portion of S_EPCn. Note that channels A, B, and C, or any combination thereof, can be the same or different. The reader 100 continues the demodulation algorithm and finally obtains the S_EPC 220 of the tags 110, 120, 130 on the shelf. The S_EPC 220 is sent to a scramble analyzer 460 that executes a scramble analysis algorithm, where the original EPC data 200 of the tags 11, 120, 130 is obtained. The EPC data 200 corresponding to each tag can be kept in the reader 100 or sent back to the original EPC manager 310 (eg, manufacturer) in the form of an inventory report. One skilled in the art will appreciate that the scramble analysis operation can be performed at other locations, such as remotely located computers or online servers. Tags 110, 120, 130 use at least a portion of the scrambled version of EPC data 220 instead of the highly structured EPC data 200 to select a channel in each path of the multi-path transmission algorithm, Collisions in the system of FIG. 4 are minimized. The scrambled data 220 is very similar to the uniform distribution data and similar EPC
Collisions between products with 200 are minimal. See Section III below for details on multi-path transmission algorithms and channel selection, and see Section V below for details on collision and collision resistance.

  No information is exchanged between the tag 110 and the reader 100 unless the tag 110 requires a channel to be used for transmission (described later). Therefore, the scramble and scramble analysis method of the present invention must be self-referential, that is, only information that needs to scramble the EPC 200 or scramble the S_EPC 220 is data.

  The system described in the present invention needs to use a scrambling method with certain important characteristics. The scramble method has an important characteristic of mapping a normal data series (such as an EPC data series) to a result showing a characteristic of uniform random distribution. In the preferred embodiment, this scrambling method has two main characteristics:

1. Suppose that k is a predetermined integer and two ordinary EPC 200s expressed in k-ary are given (for example, in a typical EPC 200 pair, many (but not all) k-arys are the same). , the probability of matching the k Decimal S_EPC 220 scrambled corresponding to those EPC 200 are n consecutive (used by the tag 110 to determine channel assignments) is approximately 1 / ^{k n.}

2. Let two regular EPCs 200 represented by k-decimals where k is a predetermined integer and the scrambled output matches for n consecutive k-decimals (used by tag 110 to determine channel assignment). Given (eg, in a typical EPC 200 pair, many (but not all) of the k-adic numbers are the same), the subsequent m k-adic numbers (tag 110 to determine subsequent channel assignments). to) coincides with a probability of about 1 / k ^m used by.

  In the EPC 200 example represented in binary, these characteristics are related to strong avalanche properties, so that each output bit depends on all input bits and single input bit changes. On average, change half of the output bits.

  In the present invention, the scramble and scramble analysis method is executed by dividing and capturing the problem using the recursive method shown in FIG. The scramble algorithm 510 receives data and length information as input and recursively scrambles the left and right portions of the data. The scramble analysis algorithm 520 executes an inverse function of the scramble algorithm 510. The operations performed for all but the base (final) level of recursion are:

For scramble method 510:
1. Split the set of data into a first part and a second part,
2.1 performing a first scrambling method on a first part of the set of data to generate a scrambled first part of the data;
3. modifying the second portion of the set of data with the scrambled first portion of the set of data to generate a modified second portion of the set of data;
4. performing a second scrambling method on the modified second portion of the set of data to generate a scrambled second portion of the set of data;
5. Modify the scrambled first part of the set of data with the scrambled second part of the set of data.

For scramble analysis method 520:
1. Split the set of data into a first part and a second part,
2.1 modifying a first portion of a set of data with a second portion of the set of data to generate a modified first portion of the set of data;
3. performing a first scramble analysis method on the second portion of the set of data to generate a scrambled second portion of the set of data;
4. modifying the scrambled second portion of the set of data with the modified first portion of the set of data to generate a modified second portion of the set of data;
5. Perform a second scrambling analysis method on the modified first portion of the set of data.

  In both the scramble 510 and scramble analysis 520 methods described above, the correction process is reversible and is selected from a group consisting of exclusive OR (XOR), remainder addition, remainder subtraction, and the like. Further, the first and second scrambling / scramble analysis methods recursively execute steps 1 to 5 until a set of data to be scrambled / scrambled reaches a predetermined length, respectively. The case must be established. In the preferred embodiment, this predetermined length is 1 byte.

  When a set of data reaches a predetermined length, a predetermined function is executed. This predetermined function is reversible and is preferably a lookup function, and in a preferred embodiment, a replacement box ("S-box") lookup function is performed. Scramble / scramble analysis of separate bits of data using a split strategy is significantly less efficient than simply using a lookup function, so it should be stopped at a single byte. Stopping at larger sizes generally makes the S-box table very large.

  There are a number of desirable properties for a given function of the present invention. First of all, this function must be reversible. The scramble algorithm 510 uses an S-box lookup function, while the scramble analysis method 520 uses the inverse of the S-box lookup function, which is reversible so that the original EPC 200 can be found. Need to be. In practice, each entry will appear exactly once because the lookup function is reversible. In the present invention, the input to the S-box is 1 byte, and the output is also 1 byte. Each S-box and its inverse function includes 256 bytes of data.

  The second characteristic of the given function is that it exhibits a strong avalanche criterion, i.e. all output bits depend on all input bits. Also, there is an additional characteristic called the strict avalanche criterion (“SAC”), and changing one input bit changes each output bit with a probability of exactly 50%. This characteristic is not strictly necessary for the present invention, but is not harmful.

Finally, the predetermined function exhibits a low dpmax. The value of dpmax is the maximum value of the entry in the XOR matrix of the lookup table. (I, j) entries in the XOR matrix are numbers in the range of 0 <a <256 such that f (a) ^ f (a ^ i) = j, where f is a lookup table. The low dpmax property allows good mixing to continue when using repeated table iterations, which is basically the case in the present invention. Such a look-up table is often retrieved by cryptography. For example, advanced encryption standard (“AES”) cryptographic algorithms use tables that meet all but SAC and work well with the present invention. By reusing this table, it is possible to save code area when AES is also used for other system functions not related to the application described in the present invention.

FIG. 6 shows a block diagram of the scramble process deployed in separate components. In the embodiment shown in FIG. 6, the input to this process is 128-bit (or 16-byte) EPC data 200. This EPC data 200 is subdivided and labeled as separate bytes P ₀ to P ₁₅ . The scramble (or scramble analysis) process starts at the top left of FIG. 6 and the two leftmost bytes (P ₁₅ and P ₁₄ ) are input to the “MixBytes” block 610. The “MixBytes” block 610 consists of S-box lookup functions 602 and 608 labeled “S” and exclusive OR operations 606 and 604 labeled “O +”. Moving down the block diagram, the next step is to take an exclusive OR 624 of the next two bytes (P ₁₃ and P ₁₂ ) and the output of the “MixBytes” block 610. The algorithm then uses the exclusive OR functions 622, 626, 628, 632, 634, 636, 642, 644, 646, 648, 652, 656, and 658 to continuously combine the data in the final stage. The functions 620, 630, 640, 650, 660, 670, and 680 are used to mix bytes until output byte 220 (C ₀ to C ₁₅ ) appears. According to the preferred embodiment, a total of 16 S-box lookup operations and 64 1-byte exclusive OR operations are required throughout this process.

  In addition to the scrambling process described above, the data 200 is encrypted (210) before applying the universal scrambling algorithm (eg, before programming the tag 110) to further ensure data security. Is possible. Various encryption algorithms (eg, AES, data encryption standards, international data encryption algorithms, etc.) are known in the technology that can be used for this task. This increased level of security is important in terms of applications where privacy is important (such as applications where tags may contain medical or financial data that should be kept confidential).

II. Power-On Scheme A block diagram of a tag 110 that falls within the preferred embodiment is shown in FIG. In an electrostatic coupling system, the antenna 701 is a pair of conductive electrodes (eg, capacitive plates), but generally may be any method of coupling energy from an electromagnetic field into a circuit. Is possible. The alternating current (“AC”) power taken from the reader 100 coupled to the tag 110 is rectified in the power converter 703 and the resulting direct current (“DC”) is used to power the tag 110, and It is also used as a tag energy monitor 704, which further enables communication (these elements will be described later). The state controller 705 operates on the tag data 220 and the communication channel selection block 240 and outputs transmission information, which is a transmission element 702 (which is well understood in the art) under the control of the channel modulator 708. Applied to load modulation elements and the like.

The data 220 stored in each tag 110 is typically stored in a low complexity (ie, low cost) circuit and responds to queries from the reader 100. Each tag 110, 120, 1
30 typically waits until the first predetermined condition is met, after which it communicates that information with a multi-path algorithm according to the present invention. The first predetermined condition is normally set to be the same for each of the tags 110, 120, and 130, but can be randomly selected or assigned in other embodiments. An example of a general flowchart showing tag transmission status is shown in FIG. In this flow chart, the second predetermined condition can be met by various metrics (eg, it is no longer possible to satisfy the first predetermined condition or the second predetermined condition is met). If)

  In a preferred embodiment, the reader 100 powers the tags 110, 120, 130 from a remote location, and the first time when the instantaneous power level of the tag 110 exceeds a predetermined threshold (generally determined by 703 and 704). 1 predetermined transmission state is reached. FIG. 9 is a flowchart of this operation, where T1 and T2 represent the first and second power level thresholds. It should be noted that one skilled in the art can employ implementations that use other predetermined conditions (such as specific sync pulses or pseudo-random pauses) without departing from the spirit of the invention. When power arrives at the tag 110 (from a remote reader 100 for passive tags or by an internal power supply for active tags), the tag 110 continuously monitors the received signal strength and transmits Determine the start time. When tag 110 begins to modulate and transmit data 250, the tag is fully activated. In the preferred embodiment of the system, the plurality of tags 110, 120, 130 are typically fully activated at a specified time.

  The fully activated tag in the group continues to send its information on multiple paths (described below) until it reaches a second predetermined transmission state, and when it reaches it, it stops data transmission. A second predetermined power level such that the received power level of the tag 110 as observed by the tag energy monitor 704 is lower than the first predetermined threshold or is usually set higher than the first predetermined threshold. The second predetermined transmission state in the preferred embodiment is reached.

  In this way, the first and second predetermined transmission states form a range of received power levels (e.g., a window) in which each group of tags is typically fully activated. In the preferred embodiment of the present invention, the power-on range is typically about 3 dB wide and the tags 110, 120, 130 are 1-2X (for some normalized received operating power level). Responds to range power. Due to this power window, the transmission of tags generally fits within a relatively narrow power window, and thus (for example, as in a spread spectrum system using non-orthogonal spreading codes). ) Normal perspective problems that affect some communication systems are mitigated.

  In the preferred embodiment, all tags 110, 120, 130 in the system are typically assigned the same power-on range, but other embodiments are possible and programmable (eg, pre-allocated but different) In some cases) or use a random power-on state. One such example may occur when different power-on range levels are assigned to different manufacturers and somehow separate (or distinguish) products from different manufacturers.

  In still other embodiments of the described system, there may be a tag with bi-directional communication capability, in which case the first and second predetermined transmission states may be some kind of sync pulse or other Consists of signal transmission information. If the predetermined transmission state is random, it can be determined randomly on the tag 110 or when the tag 110 is programmed. Again, it should be noted that other implementations of these transmission controls (eg, synchronization signals, bi-directional communication with tags, etc.) are possible without departing from the spirit of the present invention.

  In the embodiment shown in FIG. 10, the reading device 100 is a device that can be remotely controlled from the head office by a controller 1001, and is connected to an antenna 1004 installed on a shelf 1005 via a transmission medium 1003. Has been. For objects 1020, 1021, 1022 with different physical dimensions, tags 110, 120, 130 are arranged in various parts of the package, so that antenna 1004 associated with reader 100 and tags 110, 120, 130 The coupling with the associated antenna 701 will vary, and as a result, the tag electronics 1012 will receive different power levels. Due to the different coupling characteristics between the reader antenna 1004 and the various tags 110, 120, 130 in the system, different tags for a given reader antenna excitation level (ie, reader transmit power level). The received power levels are also different (indicated by range boundaries 1030 and 1031). In addition, since the various tags 110, 120, 130 begin transmitting at different reader transmission power levels and thus at different times, this effect can be used to reduce the number of tags present in the system of the preferred embodiment. It can be reduced, though roughly. However, it should be noted that in the preferred embodiment of the present invention, multiple tags 110, 120, 130 still typically begin transmitting at a specific power level simultaneously. For example, there may be 1000 items (tags) in the inventory that need to be identified, and the reader 100 will follow 10 different possible power levels, step by step, from approximately 100 tags at each power level. Can be activated (although fewer tags may be activated at upper and lower power levels in some cases). In other embodiments of the present invention, transmission from multiple tags is only possible, such as in the case of a time-slotted (channelized) system in which the user selects a specific time slot to transmit (with respect to a common criterion) ( Synchronous processing is possible (although not necessarily simultaneously). Note that in one embodiment, the reader 100 starts with the lowest transmission power level and follows all possible transmission power levels step by step. Thus, the specific power-on range of the tags 110, 120, 130 effectively controls the start and end of transmission by each group of tags. This aspect allows the reader 100 to determine when all of the tags 110, 120, 130 within a specific power-on range (eg, between 1030 and 1031) are uniquely identified, It is possible to go one step to the next power level (eg, above 1031) or end the identification process.

  In other embodiments, the reader 100 “learns” or stores a range of expected power levels for a given inventory profile, and its power level with priority given to the power level along with a history of activity. It is possible to arrange a power sweep. As described below, when the reader 100 advances one step to a power level with no activated tag, the condition is sensed (usually through short energy or modulation detection measurements) and quickly moved to the next power level. Advance one step to minimize total tag reading time.

III. Channel Selection and Transmission Methods All multiple source (or multiuser) communication methods use some kind of channelization method as in the present invention. The present invention may use any one of a plurality of channelization methods or techniques. In general, the channelization methods used in the present invention can be classified into two categories: orthogonal channelization methods or pseudo-orthogonal channelization methods.

Orthogonal communication channels have the advantage that communication on the selected channel does not (at all) interfere with communication on other channels in the linear system (ie, cross-correlation between different channels is defined as zero). Pseudo-orthogonal channels are nearly orthogonal (eg, cross-correlation values are close to 0 for different channels) and are typically used in direct-sequence code division multiple access (DS-CDMA) systems; In that case, each user is usually assigned a different spreading code.

  In the art, the cross-correlation characteristics of different phases (i.e., time shifts) of a maximum length linear feedback shift register ("LFSR") sequence (i.e., m-sequence). Is well known to be low (ie, quasi-orthogonal). The cross-correlation value for two unaligned sequences is defined as -1 / N (normalized), where N is the length of the LFSR pseudo-noise ("PN") sequence. Different code phases of the same basis m-sequence are often used to channel different users of a code division multiple access system. Each symbol or bit in the PN sequence is usually referred to as a “chip”, as is well known in the art.

  Examples of orthogonal channelization functions include Walsh functions, time slots in slotted systems, frequencies in frequency division systems, and specially augmented PN codes. Examples of the quasi-orthogonal channelization function include the m-sequence or PN sequence as described above.

  As an example of a specially augmented PN code, time-aligned (ie, synchronized) artificially inserted zeros appear on the channel with the same time offset, resulting in the same m-sequence between different code phases 2 is artificially inserted into the sequence so that the cross-correlation value of 0 becomes 0 (that is, not generated by normal operation of the LFSR). Some have a decimal value of zero. Note that in the preferred embodiment of the present invention, in a synchronized system, these particularly augmented m-sequences (how to generate are shown in FIG. 11) are used to obtain orthogonal code channels. I want. Another advantage of the spread spectrum technique used is that it also provides resistance to interference (also called processing or spreading gain), as is well known in communications technology. In a harsh electromagnetic environment such as installation in a factory, it is important to apply such a method.

  As described above, multi-path algorithms are used for data transmission by tags 110, 120, and 130 in the system being described. The multi-path transmission algorithm is important in determining the total reading time of the tags 110, 120, 130, and consists of a plurality of different aspects. The general idea adopted by this algorithm is that each tag 110, 120, 130 selects a specific (preferably uniform random) channel for communication on each algorithm path.

  In the preferred embodiment of the described system, channel selection 240 is typically based directly on data 220 stored in tag 110. Thereafter, tag 110 typically transmits information (ie, identification data) in the preferred embodiment until the next path of the algorithm on the selected channel, selects a new channel on the next path, and repeats the process. In the preferred embodiment of the present invention, it is assumed that the transmission of the tag is largely synchronized (due to the first predetermined condition).

  Channel selection by each tag is based on predetermined information (ie, typically determined by tag programming 230 or, in some cases, the design of the tag itself). In the preferred embodiment of the present invention, the channel selection of each tag 110 is determined directly (in an algorithmic manner) from the identification data 220 stored in the tag 110 (as detailed below). Furthermore, it should be noted that in other embodiments, the predetermined information may include a random number that is not directly based on the data stored in the tag 110.

  As described in detail in Section I above, in a preferred embodiment of the present invention, data 200 (eg, EPC, CRC, etc.) is stored as a key to achieve proper system performance prior to being stored (230) on tag 110. Etc.) must be randomized (or scrambled) in advance (211). The tag 110 essentially uses stored data 220, or a portion thereof (eg, 221, 222) and selects a communication channel for each path of the multi-path algorithm (240), thus improving overall system performance. It is important that the data 220 appear uniformly random in order to maximize. This is accomplished through the low complexity lossless scrambling algorithm 211 detailed in section I above.

  In particular, as shown in FIG. 12, the channel selection step 240 in each of the plurality of transmission paths of the preferred embodiment includes a predetermined value of pre-scrambled (ie, randomized and stored) data 220. This is performed by selecting a communication channel 240 in each path using a subset (eg, 221, 222, 223, 224). A channel selector 1220, such as a rectifier or multiplexing device 1240, typically selects a channel. A new subset of data 221, 222, 223, 224 (ie, subtracting a new random number) stored in tag 220 is typically used for channel selection on each subsequent path of the algorithm, and the entire multi-path transmission algorithm. It is possible to select the channels randomly and independently.

  The tag 110 can transmit all of the data 220 on each algorithm path (such as the preferred embodiment) or can transmit only a portion of the data (ie, generally used by the tag in the next path). Note that enough data is transmitted to determine the channel (s) to be Typically, the portion of data 221, 222, 223, 224 used for channel selection for each path of the algorithm is a unique contiguous section of data 220, preferably pre-randomized. The specific selection of channels for a plurality of paths in a multipath transmission algorithm is referred to as a “channel selection profile”.

For example, in a system where 128-bit pre-scrambled identification data 220 is stored in each tag, a 16 (ie, 128/8) algorithm is used using 8-bit unique but sequential sections. One of 256 (ie, 2 ⁸ ) channels in each of the paths can be selected. Thus, the first randomized byte of data for each tag (eg, 221) selects each tag's communication channel on the first path of the algorithm (240) and is randomized for each tag. The second (and preferably different) byte of data (eg, 222) is used to select a channel for transmission on the second path of the algorithm (240), and so on. This multi-transmission path process is performed until all the data stored in the tag is exhausted (for example, until the 16th path is completed in this example. In FIG. 2, this corresponds to 224), or the reader 100 Continue until a signal is sent to the tag to stop transmission (generally as sensed by the tag 110 when the second predetermined condition is met at 704 (1210) as described above). Once the data is exhausted, the entire process can optionally be repeated, but usually the tag selects the same (deterministic) channel. Note again that it is desirable to select a randomly uniquely determined channel for each algorithm path for each tag to randomize collisions that inevitably occur (see Section V below for details). See).

Of course, one of ordinary skill in the art can use other (eg, non-contiguous or not completely unique) sections of data to directly or indirectly select the communication channel for each path. . In this way, it is possible to extend the maximum number of algorithm paths before iterative channel selection without substantial limitations. It is possible to modify the channel selection profile (or channel selection algorithm) after a certain number of transmission paths and use a different subset of the same data 220 for subsequent channel selection 1220 (before pattern repetition occurs). To extend unique channel selection). For example, after 16 paths of a multi-path transmission algorithm, the channel selection data (ie, predetermined data) is shifted by 4 bits (in the example above) to arrive at a new channel selection for subsequent paths of the algorithm. Is possible. In this way, the complexity of the tag circuit is increased, but the number of unique channel selections can be increased substantially without limitation.

  In yet another embodiment of the channel selection algorithm, some kind of mapping (typically a one-to-one lookup table or other algebra or logic) function is applied and stored or programmed on the tag. It is also possible to determine channel selection from data (which is generally limited). The only characteristic that is important in the channel selection process is that the channel selection can be calculated in the reader 100 if some of the information about the data in the tag is known.

  Since the channel resources are limited (that is, the number of channels that each user can select with the multi-path communication algorithm is limited), inevitably a collision occurs on the transmitting side tag. A collision is defined as the case where two or more tags select communications on the same channel in a particular algorithm path. Such a situation is expected under normal system operation. For example, for the standard case where 25 tags communicate on 64 channels, the probability that a collision will occur at least once is 99.6% per path. This is the formula for the probability that no collision will occur when M tags are communicating on N channels:

＃１ Ｐ｛衝突なし｝
であるという事実に基づく。

# 1 P {no collision}
Based on the fact that

Some numerical examples of tag transmission collisions and countermeasures will be described in the following section V, “Collision Mitigation Methods”.
In many cases, the number of tags present in the system (at a particular power-on level) can even exceed the number of available channels (especially in the path towards the beginning of the preferred embodiment algorithm, Or if the number of available channels is set low as described below). Such a situation is completely acceptable in the present invention when orthogonal channelization means are used. Note that a normal DS-CDMA system (using quasi-orthogonal channelization code) is considered overloaded at that time and no reliable communication may occur ( Especially if there is no detailed information about the transmission characteristics of the tag). In the system described, it is important that the activated tag population can be reduced more effectively by collision mitigation techniques, which are described in detail in Section V below.

Also, the preferred embodiment of the present invention uses various numbers of channels (generally determined by 221, 222,... 224) per path of the multi-path transmission algorithm, and provides overall system performance (eg, total It is also important to improve read time, total system capacity, reliability, etc.). That is, the number of channels that can be used on one path of the multi-path transmission algorithm may be different from the number of channels that can be used on the other path of the transmission algorithm. The various number of channels per algorithm path (ie per unit time) is referred to in this description as a dynamic channel profile because the number of available channels varies dynamically with time. . Implementing a dynamic channel profile essentially optimizes the total transmission time (or total read time) of the expected tag population.

Note that the transmission time for each path of the algorithm is usually proportional to the number of channels available for that path of the algorithm (regardless of the channelization method used). The total transmission time (T _TX ) of the multi-path transmission algorithm is

で表すことが可能であり、Ｌはデータを正常に送信するために必要な伝送経路の数、Ｒは伝送（信号送受信またはチャネル・シンボル）レート、Ｂ_ｉは１パス当たり送信されるデータ・シンボルの個数、Ｎ_ｉはアルゴリズムのｉ番目のパス（または拡散利得）で利用可能なチャネルの個数である。本発明の好ましい実施形態では、Ｌは１６経路と大きく、Ｂ_ｉは１２８ビットに固定され、Ｒは６２．５ＫＨｚに等しく、特定のＮ_ｉ値は上の例で与えられた値であるが、これは、システムの１つの特定の実施形態にすぎないことに注意されたい。１経路当たりの使用可能なチャネルの個数（Ｎ_ｉ）は、以下のように（２４０にも示されているように）一般に、各経路（ｎ_ｉ）内の通信チャネルを選択するために使用されるビットの数に依存することに留意されたい。

Where L is the number of transmission paths required to successfully transmit data, R is the transmission (signal transmission / reception or channel symbol) rate, B _i is the data symbol transmitted per path , N _i is the number of channels available in the i th path (or spreading gain) of the algorithm. In the preferred embodiment of the present invention, L is as large as 16 paths, B _i is fixed at 128 bits, R is equal to 62.5 KHz, and the specific N _i value is the value given in the above example, Note that this is only one specific embodiment of the system. The number of available channels per path (N _i ) is generally used to select the communication channel in each path (n _i ) as follows (as also indicated in 240): Note that this depends on the number of bits to be used.

システムの好ましい実施形態では、Ｎ_ｉは１経路当たりの利用可能な符号位相の拡散利得および個数を表し、Ｒは毎秒チップ数単位の信号送受信レートである。高度な衝突緩和手法（以下の第Ｖ節で説明している）を適用することにより、タグ１１０、１２０、１３０からの伝送経路の必要な個数（Ｌ）を大幅に減らすことが可能であることに注意されたい。一般に、説明しているシステムの他の実施形態では、上の式の中の値に制限はない。

In a preferred embodiment of the system, N _i represents the spreading gain and the number of available code phases per path, R is the signal transmission and reception rate per second chip number of units. By applying an advanced collision mitigation technique (described in Section V below), the required number (L) of transmission paths from the tags 110, 120, and 130 can be significantly reduced. Please be careful. In general, in other embodiments of the described system, the values in the above formula are not limited.

  Since the transmission time per path is dependent on the number of channels (and symbol rate) available per path in the preferred embodiment shown in the figure, the earlier path of the multipath transmission algorithm It is possible to use fewer channels to improve the total read (ie acquisition) time performance of the system (in this case, adding more channels to the system has little benefit because there are fewer tags). Increasing the number of channels later in the algorithm (potentially multiple steps) can accommodate the case where more tags are present in the system, and the reader 100 can Do not employ the more advanced signal processing (eg, advanced collision mitigation) techniques described in Section V below.

In this way, in a system with few existing tags, the penalty of a longer transmission time for a system with a higher number of channel selections (previous) usually does not occur, and at the same time the existing tags Even in many systems, there is no significant penalty (because the earlier path of a multipath algorithm is usually quite short due to the small number of channels initially available). Also, increasing the number of channel choices for the later algorithm path ensures that a system with a large number of tags can successfully acquire all data with a limited number of algorithm paths (and therefore the reliability of the system). Increase).

  For example, in the preferred embodiment of the present invention, 128 bits of data 220 are used, the first and second algorithm paths have 32 channels, and the third to sixth algorithm paths have 64 channels. There are 128 channels on the 7th and 8th paths, and 1024 channels on the remaining 8 algorithm paths. Once again, in this embodiment, a unique subset of data 220 is used to directly select channel 1260 for history throughout each path (1220), and the unique non-overlapping portions of data are exhausted. Note that up to 16 algorithm paths are executed. In other embodiments of the invention, it is possible to use a variable number of channels per transmission algorithm path that changes after a predetermined number of paths. For example, in the first 16 paths of the example multi-path transmission algorithm above, any location from 32 to 256 available channels can be used (ie, 5 to 8 bits of channel selection data), In the next 16 paths, any location from 256 to 4096 available channels (i.e., 8 to 12 bits of channel selection data) can be used. In this way, the dynamic channel profile (or the number of available channels per algorithm path) can be expanded substantially without limitation. Again, note that it is possible to increase the maximum number of paths by driving the channel selection algorithm using overlapping or interleaved portions of data.

  The actual selection of the number of available channels per algorithm path (also referred to as a dynamic channel profile) in a particular embodiment of this system is further (in addition to the expected number of tags present in the system) ) Depending on the expected or major type of signal processing algorithm used in the reader 100 (such as the type of collision mitigation algorithm). Note that since the number of channels per algorithm path varies, the channel profile may decrease in later paths. In general, a channel profile that varies with time is considered a dynamic channel profile for the purposes of the present invention.

  It should be noted that the channel profile of a particular tag 110 does not necessarily have to be known a priori by the reader 100, but in general it may be. If the channel profile of the tag 110 is not known, the reader 100 checks the period of the PN (chipping) sequence, for example (possibly by measuring autocorrelation or spectral characteristics) and responds accordingly. Will have to operate (demodulate).

In particular, the preferred embodiment of the present invention uses random channel selection and selects a specific spreading code (or code channel at 1220) for each path of the multi-path transmission algorithm. More specifically, in the preferred embodiment, a portion of the data 220 stored / programmed in the tag 110 is used to generate a particularly increased m-sequence time offset of length N (or code like 1220). Phase) (where N is equal to the number of channels in a particular algorithm path, as described above). An outline of this process is shown in FIG. The different phases of the PN sequence generally calculate a sum modulo 2 of two or more m sequences, and output a third code phase of the same m sequence, a PN generator (LFSR) It is obtained by applying a state mask function (or AND-XOR reduction network 1100). Thus, tags 110, 120, and 130 all start from the same initial generator state within each algorithm path so that all of the transmissions of tags 110, 120, and 130 are synchronized to a known basic initial generator state. Use the same basic LFSR (m-sequence) generator. These aspects are important for quick and effective demodulation in the reader 100, as described in Section IV below. Note that the length of the basic LFSR sequence generator (i.e., the primitive polynomial) typically changes dynamically (the number of channels changes) for each algorithm path, as described above.

  Conventional m-sequence generators usually become a specially augmented PN sequence generator by forcing a zero output for the first chip (or PN bit) time in the preferred embodiment, and sequences from different tags Is always zero over a given sequence period. In addition, if 0 is forcibly entered into the code sequence at another time point, an orthogonal sequence is output (similar to other systems). In other embodiments, another type of orthogonal function generator is replaced with an LFSR PN generator ( Note that it can also be used instead of, for example, Walsh or Hadamard functions). Thereafter, the data 220 stored in the tag 110 is spread by the generated spreading code 1260 using conventional means 1230 (eg, XOR in a digital implementation, as is well known to those skilled in the art). Gate, or divider in analog implementation). The activated tag spread data signal is then transmitted (collectively) on a given communication channel.

  Note that a tag can also use a range of modulation types for data transmission (eg, amplitude modulation, phase modulation, frequency modulation, or combinations thereof). Although the preferred embodiment of the present invention uses some form of amplitude shift keying (“ASK”) from load modulation via transmission element 702, other modulation types and implementations are certainly possible (eg, difference Dynamic quadrature phase shift keying, quadrature amplitude modulation, pulse code modulation, pulse amplitude modulation, pulse position modulation, etc.). Various data encoding and mapping techniques can also be used in the present invention. Examples of encoding techniques are zero return (RZ), non-zero return (NRZ), Manchester, and differential encoding, all well known in the art. It should be noted that various types of encoding, modulation, code and signal transmission / reception can be used in the present invention without loss of generality, as is well known to those skilled in the art. Examples of coding techniques include CRC codes, convolutional codes, block codes, etc., all of which are well known in the art.

  Also, since the preferred embodiment tags 110, 120, 130 directly modulate the carrier wave provided by the reader 100 via the transmission element 702, there is no local oscillator (although using a locally oscillating carrier is also reliable. In addition, it is possible within the scope of the present invention and does not limit the application in any way). In a preferred embodiment of the present invention, the power converter 703 rectifies the carrier signal from the reader 100 and supplies power to the circuitry on the tag 110 from where the reader 100 is remote. It should be noted that actively powered tags can be used and do not limit the use of the invention in any way. The general goal of the present invention is to minimize the complexity of the tag 110, and the circuit of the tag 110 can be kept to a minimum by using the described approach of the preferred embodiment. is there.

IV. Fast Demodulation Scheme As shown in FIG. 13, the reader 100 typically initializes the output of the signal source 1310 by means of a transmission level controller 1320 and an amplifier 1330 and transmits some minimum level of power, thereby enabling the tag 110. , 120 and 130 are started. In the preferred embodiment, reader 100 then begins transmitting that level of continuous waves. When the reader 100 begins to transmit at a particular power level, it typically monitors the signals returning from the tags 110, 120, 130 (by the coupling device 1340 and the antenna 1345). This activity detection may take the form of modulation or energy detection measurements, such as detecting the signal level or signal amplitude on each of the possible communication channels (detailed below). This measurement period is preferably as short as possible, and if no tag is activated at a particular power level, the reader 100 can quickly advance one step to the next power level ( In general, in an increasing direction). If a signal is sensed at a particular transmit power level, the reader 100 may begin a full demodulation process 1390 (possibly employing a collision mitigation technique as described in Section V below). ). In other embodiments of the system, without losing generality, the reader 100 can also send a modulated carrier signal, a sync pulse, or an asynchronous carrier waveform.

  Signal processing performed by the reader 100 may be performed by hardware or software architecture, or a combination thereof. Exemplary embodiments include some selectivity 1365, amplification 1370, analog to digital conversion 1375, and DC acquisition and gain control function 1380. In general, the reader 100 cycles through each possible communication channel within a given transmission path (sequentially or simultaneously) and searches for signal energy for each. As is well known to those skilled in the art, there are many methods available for detecting the presence of a signal and detecting the presence or absence of a collision, depending on the type of modulation and signal transmission / reception. In general, the reader 100 may also perform active or passive suppression 1360 of the carrier signal to provide interference or noise cancellation in some embodiments (sources other than the desired tag in the system). About any form of interference from).

  As mentioned above, the preferred embodiment of the present invention uses spread spectrum modulation in the tags 110, 120, 130. Thus, as is well known in the art, the received data must first be despread in the reader 100 for the code channel written by applying each possible spreading code.

  More specifically, the preferred embodiment of the present invention uses a specially augmented m-sequence as the spreading sequence in tag 110, so that it is very fast, efficient and novel demodulation (ie, , Despreading and channelization) techniques can be used in the reader demodulation process 1390. Using these techniques substantially reduces the processing power required within the reader demodulation process 1390 (eg, about 57 times in the preferred embodiment), resulting in a reduction in reading time of the reader 100, Mounting costs are also reduced. The actual processing savings depend on the number of channels used in each path of the multipath system, which is the number of conventional despread operations and the improved despread operation per symbol. Can be expressed in terms of a factor (F) equal to the ratio to the number of times (using a novel combination of reordering of received sequences and fast Hadamard transform (FHT)),

Ｌは、送信元データを正常に復調するために必要な経路の数に等しく、Ｎ_ｉは（ここでもまた）ｉ番目の経路内のチャネルの数に等しい。この係数は、直接、読取装置復調処理１３９０内の処理の節約程度（通常は、毎秒１００万回オペレーション（ＭＯＰＳ）または毎秒１００万命令（ＭＩＰＳ）で表される）を表す。したがって、この例では、５７倍能力が低い（例えば、１０ＭＯＰＳ対５７０ＭＯＰＳの）プロセッサ１３９０を、最良のケースの好ましい実施形態の読取装置１００で使用することが可能である（後述のように衝突緩和はない）。

L is transmitted equal to the number of pathways necessary to demodulate correctly the original data, N _i is equal to the number of channels in (again or) i th path. This factor directly represents the degree of processing savings in the reader demodulation process 1390 (typically expressed in 1 million operations per second (MOPS) or 1 million instructions per second (MIPS)). Thus, in this example, a processor 1390 that is 57 times less capable (eg, 10 MOPS vs. 570 MOPS) can be used in the reader 100 of the best case preferred embodiment (collision mitigation is described below). Absent).

  The specially augmented m-sequence (shown in box 1120 in FIG. 11) is an orthogonal extension of the conventional PN sequence and has some similarities to the orthogonal Walsh code (shown in box 1420 in FIG. 14). Note that the two sets of sequences contain the same number of binary numbers 1 and 0 in the sequence. In fact, the two types of sequences (ie, the specially augmented m-sequence and Walsh sequence) are related using a single special reordering function. This special reordering function is derived directly from the primitive polynomial used to generate the basis m-sequence in reader reader block 1520 of FIG. 15 (as indicated by tag sequence generator 1110). . By using the sequence order changing function 1510, the order of the data samples (or elements) is directly changed when the receiving device 1375 receives the data samples (or elements). Receiving device 1375 may be an analog to digital converter, an analog sample and hold device, a register, or other device that receives signals. Note that the single sequence reordering 1510 function applies to composite received signals consisting of transmissions from multiple tags 110, 120, 130 using multiple composite channels (or code phase as 110). I want.

  After the order of the received sequence is changed in a storage medium such as memory buffer 1530, it is similar to a sequence from a set of valid Walsh sequences, and is a fast transform technique such as fast Hadamard transform (FHT). Can be used to quickly (and simultaneously) despread data (as indicated at 1540) from the tag 110 for all data channels. As is well known in the art, FHT is used to quickly (in parallel) determine the correlation of data sequences for a set of Walsh codes. Without departing from the spirit of the present invention, it can be used with fast correlation determination methods that describe transformations related to FHT (eg, fast Walsh transform, Walsh-Hadamard transform, recursive Walsh transform, etc.). It should also be noted that all the processing techniques described can be implemented in the field of analog or digital signal processing.

There is plenty of literature on conventional FHT algorithms (eg, as shown in box 1410), and basic kernel operations (box 1400, labeled “Butterfly”) are shown in FIG. Note that The radix-2 FHT butterfly is similar to the radix-2 FFT butterfly, which consists of multiplying (or equivalently adding and subtracting data values) the data elements and +1 and -1 values. An 8 × 8 FHT lattice structure 1410 is also shown in the figure. Each output of FHT 1550 is referred to as an FHT bin or FHT code channel. Since FHT is a fast transform, it proves that the processing savings over traditional correlation (similar to the coefficient F represented above) are equal to (N ² / N log N) for N-point orthogonal sequences. It is possible. This same saving is achieved using the fast correlation determination technique described.

The exact received data reordering function 1520 is determined by observing the tag's Fibonacci LFSR (shown at 1110) cycling during normal operation (see also example below). In order for the LFSR to progress, incoming (spread) received data samples must be stored in the received data memory buffer (1530 or other storage medium) when received in time (linearly). Supports indirect addresses directly. This sequence of addresses (in 1520) is not actively generated in the receiver, but alternately on the storage medium (eg, random access memory, read only memory, hard disk drive, etc.) Can be stored. In this way, the order of the received m-sequence (or the sum of the m-sequence) is changed so that the elements (or more specifically, the rows in the Hadamard matrix) in the Walsh sequence are accurately represented. Thus, it is possible to despread the received data channels in parallel using conventional Hadamard transform (correlation) methods (at 1540). Note that this data sequence can also be double buffered in memory to match the processing latency.

  The output index (or bin) 1550 of the FHT indicating the signal energy directly corresponds to the mask value 1130 (when expressed in binary) used in the AND-XOR reduction 1100 of the tags 110, 120, and 130. For example, the channel selection code 1130 (“c0-c4” shown in FIG. 11) (transmitter processing) directly corresponds to the active output 1550 (receiver processing) of the FHT block 1540 of FIG. Note that a binary mask value 1130 is applied within the tag 110 to select a particular code channel (or code phase). This is also shown in FIG. 7 where mask 710 is drawn from tag data 240 to channel selection 240. That is, the binary mask value 1130 (and the FHT bin index) is directly in the data 221, 222, 223, 224 stored in the tag 110 used to select the channel in a particular path. (See also identifiers 1710, 1820, 1830, and 1840 in FIGS. 17 and 18 for a supplementary explanation of the relationship between tag data and channel selection). Each tag 110 transmits its data 220 on a fixed channel 1260 during the duration of each pass of the preferred embodiment multi-pass algorithm. The output signal level of each FHT bin directly corresponds to the signal level on each code channel 1260 after despreading (eg, for each code phase). As detailed below, the data signal 1550 at the output of each active FHT bin in the channel selection portion of the received data sequence can be verified by checking against the binary FHT index value (because two This is because the data series should be checked against valid data). This technique is illustrated in FIG. 18 for path # 2 of the multi-path transmission algorithm. Note that the data series 1820, 1830, 1840 on the portion 222 used to select the channel 240 for the second path is a binary value corresponding to the FHT bin number.

  By using the combined reordering and FHT approach shown in FIG. 15, this demodulator provides fast demodulation (ie, despreading) of all possible code channels (ie, code phases) of the preferred embodiment. Is possible. An N-point FHT is typically required to demodulate the N channel for each received symbol period in the receiver (corresponding to the required dechannelization and despreading operations for each potential data channel and symbol). Note that there are. Also, in other embodiments of the transponder system, it is possible to use orthogonal Walsh codes for the channelization function, in which case the FHT bin corresponds directly to the Walsh code channel index (and the reordering process is Not necessary). However, since Walsh channelization codes are periodic and can be highly correlated with periodic interference sources, such systems do not have as good interference cancellation capability when compared to the preferred embodiment. Thus, the preferred embodiment of this system uses a specially increased m-sequence as the channelization function and uses the demodulation technique described above.

For example, in a system using a specially increased PN sequence of length 16 (N = 16, N = 4) in the tag transmitter, the channel selection value 1130 (n _i ) of binary “0001” (1) The sequence 1260 represented by is “0111101011001000”, but the sequence 1260 represented by the channel selection (mask) 1130 of the binary number “1001” (9) is “0010110010001111”. It is only a different time shift or code phase of the same basic m-sequence that is specially augmented). FIG. 11 shows an example of the tag PN generation and mask circuit of the primitive polynomial of 23 (when expressed in standard octal notation). The two tag transmitters are assumed to transmit these sequences independently over the communication channel. The reader receiver resolves these two signals using a special reorder function 1520 and FHT processing (as shown in FIG. 15). The special received data sample order change that must be used for the transmitted PN sequence is in this example {0, 15, 7, 11, 5, 10, 13, 6, 3, 9, 4, 2 , 1,8,12
, 14, 1120}. This sequence replicates the m-sequence generator 1110 used in the tag 110 and observes the state of the PN generator or simply stores the necessary reordering sequence in memory within the reader 100. It is possible to generate. The incoming received data sample stream is stored in memory by indirect addressing using a reordering sequence. For example, the first valid A / D sample that arrives at the reader is stored in memory buffer location 0 of storage medium 1530 (as is the case with all specially augmented codes) and the second sample. Is stored in memory location 15, the third sample is stored in location 7, and so on. When N (16 in this example) samples are received, normal FHT processing 1540 can be performed on the newly reordered data samples in memory buffer 1530. The order change function converts the “0001” PN code to the sequence “01010101010101101” (same as Walsh code 1), and converts the “1001” PN code to the sequence “01010101110101010” (same as Walsh code 9). FHT 1540 indicates that signal energy is present in bin 1 (corresponding to channel code 1) and bin 9 (corresponding to channel code 9) of output 1550 (eg, tag is being transmitted). . Therefore, the remaining tag data can be detected by observing the FHT output of bin 1 and bin 9 for each transmission symbol.

  In the above approach, the first chip (or symbol) transmitted by the tag 110 is binary 0 (+1 normalized signal on the channel) even if no such signal is actually transmitted in the receiver. Note that it can be used for conventional (ie, not specifically augmented) m-sequences by assuming that it is equal to the value. Accordingly, the first buffer location in storage medium 1530 is initialized to a +1 value and processing (ie, reorder 1510 and FHT 1540) continues as usual. In this way, it is possible to obtain a correlation between a conventional PN sequence and a plurality of code channels (or code phases) at a very high speed. Other normally augmented PN sequences can also be accommodated by tracking where additional chips (eg, other than the first chip above) are inserted into the sequence.

  The fast correlation determination techniques described above (ie, specific received sequence reordering 1510 and FHT 1540) can be generated by an AND-XOR reduction network 1100 (whether or not generated by such a network). And apply to communication systems using Many popular communication systems use these types of PN sequences or sequences generated from a combination of conventional m-sequences (such as Gold codes, as is well known in the art). Examples of such systems are IS-95, IS-2000, 3GPP CDMA mobile phone systems, and GPS CDMA positioning systems. The fast correlation determination techniques described above may be equally effective within those systems.

  In any case (regardless of the removal technique employed), the composite received signal is filtered and amplified in the receiver front end 1610 and then in the reader 100 as shown in FIG. Must be channelized (or dechannelized) 1620. Each channel is then generally processed separately (but in some cases simultaneously) for signal and collision detection (generally 1630). For example, in another embodiment of a system that uses a Walsh code instead of the m-sequence described, it is also possible to demodulate all different data channels simultaneously using FHT operations as is, as described above. is there. In another embodiment of the system, a conventional bank of despreaders (in parallel or time division) (instead of 1540, 1620) may be used to perform the dechannelization and despreading steps. Is possible. As is well known in the art, a despreader is typically an integrator and dump function placed after a multiplier.

  In other embodiments, in other examples of communication systems, orthogonal time slots can be used as channels (such as in slotted ALOHA systems), in which case signals from different tags (separate times) To) and demodulated. Note that the selected channelization method does not change the general types of collision mitigation algorithms that can be employed within the reader 100, as described below.

  Also, since it is not possible for every tag to normally transmit its information normally in the first path of a multi-path transmission algorithm, in many embodiments of the present invention, the demodulation process is generally a multiple iteration process. Note that Thus, the reader 100 continues to receive incoming data until all data from the tag is successfully received (further using the method described below) and remains powered on (at the same power level). Must be demodulated. In addition, when advanced collision mitigation techniques 1630 are used in reader 100 (as described below), multiple demodulation iterations (eg, FHT) may be required for each pass of the multi-pass algorithm. . It is also noted that in subsequent passes of the multipath transmission algorithm, it may be necessary to tune the demodulator to the newly set number of channels, as explained in the description of the dynamic channel profile above. I want.

V. Collision Mitigation Method As described above, the number of communication channels that can be used by the tags 110, 120, 130 to communicate with the reader 100 is limited within this (and any) communication system. Because the number of communication channels is limited, and there are no organized assignments between multiple tags (that is, random assignments are actually used), the system described is sending from tags It is inevitable that a collision will occur. A collision is defined as an event or event when two or more tags choose to communicate on the same channel simultaneously (ie, on a particular path of a multi-path transmission algorithm). Note that the assignment is actually random, as shown in Section I of this specification, because the data stored in the tag is very close to uniform random data.

  In the system described, the collision mitigation technique of the reader 100 may or may not be used (as described below) depending on the complexity desired by the reader 100. For example, an inexpensive receiver may not use the collision mitigation technique, but an expensive (high processing capacity) receiver may use an advanced collision mitigation technique.

  In the following outline explanation, first, it is assumed that a specific collision mitigation technique is not used, and later, a case where the collision mitigation technique is used will be examined. Note that tags 110, 120, 130 generally transmit the same pattern regardless of whether collision mitigation is used in reader 100 or not. Each tag (eg, 110) is substantially unrecognizable from other tags (eg, 120, 130) present in the system. When the following additional steps are performed, a demodulation process in the receiver is further performed.

  In particular, in the preferred embodiment, the presence of an ASK signal with a small deviation from the tag typically takes the normalized signal by subtracting the average signal level (ie, the DC value as in 1380) from the channel. , By examining the absolute value of the remaining (normalized) signal. Note also that some form of automatic gain control (also in 1380) can be further applied to normalize the signal level. If the absolute value of the normalized signal level exceeds a threshold (generally over a period of time), the signal is said to be on that channel of the preferred embodiment.

When a signal is detected on a particular channel, the reader 100 must usually detect that a collision has occurred on that channel. This is usually detectable by examining the variance of the absolute value of the normalized signal level over a period of time. If the variance of the signal's absolute value exceeds a (different) threshold, the collision is said to have occurred on that particular channel (because the binary data values of the ID data of different tags are colliding-FIG. 17 Otherwise, a single signal is said to be on that channel (see FIG. 18). Again, those skilled in the art will appreciate that these measurements can be filtered or averaged and indexed to increase their reliability (eg, estimated values). Increase SNR). Thus, the longer the observation period for such a measurement (used in later filtering), the greater the accuracy and reliability of the estimate. Collisions can also be detected by other means such as standard error detection (eg, CRC) means, but those methods may not detect the collisions correctly in all cases. Note that if a collision occurs on a certain channel, it is possible to use standard error correction means to correct transmission errors and increase the accuracy of the signal estimate. These measurements are typically performed on all available (possible) communication channels in a given path (this depends on the number of paths in the multi-pass algorithm, as described above).

  Thus, the reader 100 typically has a signal present on all (and each) of the possible communication channels per path, and whether a collision has occurred on each channel when a signal is present. Characterize. Note that a collision is generally defined as the case where two or more tags use the same communication channel in the same path of a multi-pass algorithm. If a collision occurs on a given channel, data for that channel will generally be lost unless a collision mitigation technique is used. If a signal is present on a given channel and no collision is detected, the particular signal on that (given) channel is said to have been successfully received, and the reader 100 will generally recognize that particular Recognize the entire tag data series.

  In some embodiments, it is further possible to perform error detection or correction (or some other type of signal integrity measurement) to confirm that the data is valid and received correctly. Please be careful. It should also be noted that if tag channel selection data is transmitted, the reader 100 can also check that the tag 110 has actually communicated on the expected communication channel. (Used as another form of error checking for the portion of data used to determine the channel as described above-see also FIG. 18. Channel selection data 222 for the second path is 1820, 1830. , Must match the channel selection, as identified at 1840).

  When the signal from tag 110 is recognized (and possibly verified), it is ignored or removed from the rest of the signal population (described below). Some collision mitigation is implemented when the signal from a particular tag is actually removed from the signal population (using various possible algorithms). Note that signal removal may not be accurate to take advantage of collision mitigation.

  FIG. 19 shows a general flow diagram for the operation of the reader when using the collision mitigation technique. In this case, the reader 100 attempts to resolve as many collisions as possible (eg, errors in the data) before moving on to the next path of the multi-pass transmission algorithm (eg, in the preferred embodiment, the reader's By keeping the transmission power constant).

As described above, the reader 100 generally keeps transmissions at a specified power level until some confidence level (or probability) is obtained in which all active transmission tags have been identified.
If the signal is not actively removed (subtracted) from the signal population (or composite received signal), it is said that no collision mitigation has occurred. In that case, all data from the tag can be normally acquired (or demodulated) by using various algorithms in the reader 100. The general idea in this case is that each tag selects a communication channel that is unique (ie, occupied by a single user) in at least one of the multiple paths of the multipath source device transmission algorithm. Wait for. This technique is generally the lowest complexity identification method available within the reader 100, but is generally also the slowest (ie, the longest total transmission time for exchanging information).

  One very low-complexity algorithm for the case where the collision mitigation technique is not used by the reader 100 is a simple method in which the tags 110, 120, and 130 transmit the maximum number of paths using a multi-path communication algorithm. Take. The maximum number of paths is typically determined (as described above) when the unique portion of the data stored in the tag is exhausted.

  As described above, the reading device 100 directly controls the number of paths when the tag transmits by controlling the first and second predetermined transmission states. In a preferred embodiment of the present invention, the reader transmit power level is kept constant to continue transmission between fully activated tags, but the other first and second predetermined transmission states. Can control transmission from a tag in units of groups. The maximum number of paths is generally determined by a particular channel selection algorithm, but generally the data length (number of bits) for a completely unique (non-overlapping) channel selection item is the channel selection portion (number of bits) of the data. ) Is divided by the sum of). Thus, in the example shown above with 128 bits of data and 8 bits of channel ID selection data in each path, a maximum of 16 (ie, before the channel selection begins to repeat) (ie, 128/8) communication path. Thus, given the channel (eg, PN) symbol rate in the preferred embodiment, the maximum interrogation time can be determined, and the total acquisition (or read) time is in all cases Is fixed (as also shown in the above formula).

  Other (often more complex) algorithms that do not use collision mitigation techniques are possible. In one such alternative approach, a limited number (maximum number) of tags 110, 120, 130 is obtained so that a specified confidence level is obtained that the received data (or acquired tag inventory) is correct. Less than) path. This is typically the expected number of source devices (or tags) present in the system (or at each power-on level) and the desired confidence level (or probability of successfully identifying a product or tag in the system) Determined by. For example, from the simulation using the dynamic channel profile given in the example above (more than 1000 trials), it turns out that an average of 7.73 transmission paths are needed to identify 50 tags. However, in order to uniquely identify a tag after 1000 trials, a maximum of 10 paths are required. Thus, the reader 100 is sufficient to keep the power on at a given power level for 10 passes and that all 50 (about) tags have successfully transmitted their data on a unique channel. There can be reliability. Again, reader 100 need only be able to determine when there is only one tag 110 on the channel to receive the ID data. As a result, only 10 passes are executed instead of the absolute maximum of 16 passes given in the above example, so that the total acquisition time is substantially shortened. Detailed simulation, statistical or probabilistic analysis could be applied to determine the number of paths for other confidence levels or a given number of tags. In some applications, the reader 100 utilizes the maximum number of passes when initially inventoried and then counts the number of tags expected (or measured or observed) to be present in the system. Note that it is possible to adjust the number of passes based on.

  Apart from that, the algorithm used by the reader 100 tracks the expected collision location (ie channel) of each tag (after its data or ID information has been successfully received) and is still identified in the system. It is also possible to estimate how many unused tags remain. Therefore, the reader 100 may be able to stop the interrogation process earlier than the above method (after it is determined that there is no possibility that another tag exists in the system). In other words, the necessary number of transmission paths is not calculated in advance based on the expected number of tags as described above, but is estimated by the reading device 100 while watching the state at the time of reception. This technique is described in detail in the following example and FIG.

  In more advanced embodiments of the reader 100, any one of several forms of collision mitigation techniques can be used. Using collision mitigation techniques generally eliminates the effects of collisions on a given communication channel. Ideally, the effects of certain collisions on the channel are eliminated. This can be achieved in the desired system by reproducing (at least conceptually) a known signal and subtracting the known signal from the total signal population (or composite received signal). . Note that this interference signal cancellation can be done at any stage of the demodulation process (eg, it can be done at the chip rate or after despreading in the preferred embodiment). In a preferred embodiment of the invention, collision mitigation is performed after demodulation (ie, despreading) to reduce implementation complexity.

  In general, when using collision mitigation techniques, the more signals that are known, the fewer tags that appear to be present in the system for the specified path of the multi-pass algorithm. In the preferred embodiment, the data stored in the tag 110 directly determines the channel selection (or is known by the reader 100 by some other means) so that the reader 100 has successfully received the data. Later (generally when tag 110 transmits on some other unoccupied channel), tag 110 recognizes all channel selections performed for all paths of the multi-path communication algorithm To do. Therefore, the reader 100 can predict a channel that the tag 110 uses for future (and past) transmission as described above. The observed signal level from the tag 110 is also typically measured in the reader 100 (and passed through a low-pass filter) in a normal signal detection process and provided (non-collised) tags. A reliable estimate of the actual signal strength of is obtained. This knowledge can be used to effectively regenerate known signals and accurately subtract them from the entire received signal, thereby removing the effect from other transmission paths. It is.

  In general, there is a group of collision mitigation techniques at various levels of complexity, which are generally more complex than implementations that do not use collision mitigation techniques (eg, require more processing power, memory capacity, or hardware). To do). However, when such a method is adopted, generally, the total tag data acquisition (reading) time is considerably shortened, and the system capability can be greatly increased. Again, the channel is assumed to be quasi-static and the system is relatively linear, providing the best system performance.

A relatively simple form of collision mitigation requires subtracting a known signal (forward in time) from subsequent passes of the multi-pass algorithm. For this reason, this form of collision mitigation is generally referred to as forward collision mitigation. FIG. 20 shows an example of a flow diagram for a reader process that uses a forward collision mitigation technique, where the process is performed sequentially (eg, one channel at a time) for clarity of the process. To do. This process generally determines which tag 110, 120, or 130 has successfully transmitted the ID data (as described in the receiver algorithm above) and finds out for each pass of the multi-pass algorithm It is necessary to maintain a data structure (or list) that contains the channel selection (tag) and the estimated signal level. Once identified, the tag ID data and the signal level of the transmitted tag signal can be effectively removed from subsequent collisions with that tag. Again, it should be noted that it is possible to increase the accuracy of the interference signal level by measuring the signal level and lengthening the filtering time. Thus, in one embodiment of the present invention, the tag signal is estimated (determined within a certain level of accuracy) and then subtracted from the appropriate (predetermined) channel in the later path of the multipath transmission algorithm. This cancels out the interference effects of the (known) tag signal on other signals transmitted by other users. Typically, this approach is possible because of the deterministic nature of each tag's channel selection based on the data stored in the tag 110.

  Since the measured signal level is assumed to be preserved in all subsequent paths, the pseudo-channel assumption is important here. In general, it is also possible to update the signal level estimate for each transmission path in order to accommodate a slow change in channel conditions. To execute this algorithm, it is only necessary to store the known tag signal information (usually contained in the data structure or list) and the composite received signal from the current transmission path (or burst). Note (as opposed to storing all received bursts in memory as described in the algorithm below). In general, using this type of forward collision mitigation algorithm may significantly improve (2-4 times) the total read time compared to methods that do not perform collision mitigation.

  Another more advanced form of collision mitigation involves subtracting known signals from both subsequent and previous paths of multipath transmission. This is possible because after the data from the tag 110 is identified, it is possible to check the channel that the data occupied in the previous pass and invalidate the involvement in the previous collision. It is. This class of collision mitigation algorithms is commonly referred to as a bi-directional collision mitigation technique. Bidirectional collision mitigation performs complex calculations (generally requires more memory to store the pre-communication path), but as a result, all tag read times are significantly reduced (no collision mitigation is performed) Compared to the method, it is shortened by about one digit).

  In general, this method requires storing a data structure that contains the channel selection and estimated signal level known in each communication path (as described above) for the identified tag. However, since the signal is subtracted from the previous transmission path (in addition to the current path as in the forward collision mitigation algorithm), further collisions can be resolved. For example, if the data from the third path of the multi-pass communication algorithm is resolved (ie, received successfully), the data from other users is the path before the algorithm (eg, the second path) In addition, this allows other users who have already collided in either the previous path (eg, the first path) or the subsequent path (eg, the third path) to be transmitted. It is possible to release. Each time data from a new user is resolved, the reconstructed signal is subtracted from all transmission paths (up to the current path) and the number of collisions is again evaluated (for all possible communication channels). In this way, the reader 100 circulates all available transmission paths (including the current path up to the current path), and there are more users who can resolve any path (up to the current path). More tag signals can be resolved substantially continuously until the point at which they run out is reached. The reader 100 then proceeds one step to the next power level and continues the bidirectional collision mitigation algorithm. This effect can be very powerful in later transmission paths, and can solve significantly more tag signals than the available number of communication channels.

Once all of the tag data has been received, the reader 100 can check the integrity of the data via the means described above (eg, error detection and correction). Reader 1
00 can also perform post-processing of the data, and typically includes functions such as scramble analysis, decryption, classification, and removal of redundant items (in the preferred embodiment of the present invention, multiple power-on Power on in range). Note that some or all of these functions can also be centrally located, thereby handling multiple readers or antennas.

Examples of system operations The operations of these algorithms are best understood using examples. These examples detail a simplified hypothetical system of tags that draw random channels per path. 21, 23, and 24 are system state diagrams used to illustrate this example, showing the channel that each tag chooses to communicate for each subsequent pass through the algorithm. The state shown in the example is the original output of an actual experiment that uses a random number generator to select a channel. The type of physical channel (eg, code phase, time slot, etc.) is irrelevant at this point. Thus, as detailed in section I above, there is a data scrambled portion of the present invention, so a relatively accurate model of the entire system must be prepared.

  The examples detailed in FIGS. 21, 23, and 24 assume a population of 8 tags, and the fixed channels per path of the 8 channels from which the tags are drawn for communication. Assume size. Thus, in the preferred embodiment, 3 bits (in a unique subset) of each tag's ID information are used to select one of the 8 channels used by each tag 110 for transmission on each path of transmission. select. Using the octal number, the first 30 bits of the tag ID are randomly generated, but for convenience, the following is repeated.

Tag 1: 0033 0436 07. . .
Tag 2: 1106 2551 65. . .
Tag 3: 4767 4416 . .
Tag 4: 2044 6111 36. . .
Tag 5: 6072 3355 74. . .
Tag 6: 1476 5432 . .
Tag 7: 5443 3675 34. . .
Tag 8: 2135 5115 64. . .
Tag 1 continues to select channel 0 on path # 1, channel 0 on path # 2, channel 3 on path # 3, and so on. Tag 2 continues to select channel 1 on path # 1, channel 1 on path # 2, channel 0 on path # 3, and so on. From this list, for path # 1, which subtracts the channel from the first octal number, tag 1 is the sole occupant of channel 0, tag 3 is the sole occupant of channel 4, and tag 5 is the sole occupant of channel 6. It can be seen that the tag 7 is a single occupant of the channel 5. Since these channels have no collisions, tags 1, 3, 5, and 7 have been successfully identified as a whole, and tags 1, 3, 5, and 7 have their full IDs in the channels that had no collisions. Communicated. However, in path # 1, tags 2 and 6 collide on channel 1 and tags 4 and 8 collide on channel 2. These tags cannot be successfully identified and need to resolve subsequent passes. The reader 100 observes that a collision exists and leaves the applied power at the current level so that all tags draw another channel from the second octal number in path # 2. To be possible. Note that the tag has no way of knowing if the ID could be successfully transmitted.

In pass # 2, tag 3 is the only tag not involved in the collision. Since this tag has already been identified by the path # 1, the reading apparatus 100 did not obtain new information. None of the tags that collided in pass # 1 are still possible to identify. Statistically, 8 tags and 8
In one channel, the probability of at least one collision occurring is 1-8! / 8 ⁸ = 99.76%. This result is a more general case of the probability of no collision between the M tags on the N channels given above.

＃１ Ｐ｛衝突なし｝
とＰ｛衝突｝＝１−Ｐ｛衝突なし｝という事実から得られる。この同じ確率は、アルゴリズムを通るパス毎に衝突が少なくとも１回あるという確率である。タグとチャネルのこのような組み合わせについて、実験を１０００００回行って平均をとると、８個のチャネルのうち２．７４９８個は１パス当たり未占有であり、これらのチャネルのうち３．１３８６個は、単一タグを含み、１．５７３７個のチャネルは２つのタグを含み、０．４４８２個のチャネルは３つのタグを含み、０．０７９６個のチャネルは４つのタグを含み、０．００９３個のチャネルは５つのタグを含み、７．２×１０^−４個のチャネルは６つのタグを含み、４×１０^−５個のチャネルは７つのタグを含み、１チャネル内に８つのタグというケースは記録されていなかった。

# 1 P {no collision}
And P {collision} = 1-P {no collision}. This same probability is the probability that there will be at least one collision per path through the algorithm. For such a tag / channel combination, taking an average of 100,000 experiments, 2.7498 of the 8 channels are unoccupied per path, and 3.1386 of these channels are unoccupied. , Including a single tag, 1.5737 channels including 2 tags, 0.4482 channels including 3 tags, 0.0796 channels including 4 tags, 0.0093 Channel contains 5 tags, 7.2 × 10 ⁻⁴ channels contain 6 tags, 4 × 10 ⁻⁵ channels contain 7 tags, 8 tags in 1 channel Was not recorded.

Example without collision mitigation Without collision mitigation, the tags must all appear in the channel to be identified. This happens if the experiment is allowed to run a sufficient number of times. However, since the number of bits in the tag ID 220 information is limited, the experiment can only be performed a limited number of times before starting the iteration. For example, if the tag ID is 96 bits long and the channel is drawn using 3 bits per path (1 out of 8), the process is repeated after 32 experiments. Due to the high probability that there is at least one collision per pass (99.76% in this scenario), the ID of the tag may be “hidden” in the collision on a per-pass basis and across all paths through experiments. The probability of getting is small but a finite value. This does not mean that tag ID 220 is the same as the ID of a different tag throughout (not allowed by the assumption that there is a unique tag ID and a unique reversible mapping to a scrambled tag ID. ). This means that the tag ID 220 is the same as the ID of at least one other tag when examined with respect to the small number of bits (in this case 3) used to define the channel space for that path. I'm just saying. This introduces the concept of inventory or product uncertainty, which is only known with certain reliability of the tag inventory.

  In the experimental example of FIG. 21, eight trials are required for each tag in order to appear in a channel without collision. As already mentioned, tags 1, 3, 5, and 7 are identified with path # 1, tag 2 appears with path # 3, tags 4 and 8 are identified with path # 4, and tag 6 is identified with path #. Does not appear until 8. Tag 6 is a good example of how a unique tag can be hidden in a collision even though it has a unique ID. If this experiment was only performed in pass # 7 (ie, if the ID length was only 21 bits), tag 6 would not have been identified.

In pass # 1, four tags are identified. Two more collisions are identified, indicating that there are at least four other tags (since at least two tags cause a single collision, at least four tags are required for two collisions to occur Is). Thus, after the first pass, the reader 100 can determine that there are four known tags and at least four unknown tags, or at least eight tags in all.

  In path # 2, only a single known tag occupies a unique (unused) channel. Since reader 100 knows the complete ID of tags 1, 3, 5, and 7, it knows the channel that the tag occupies on the next and all subsequent passes. The reader 100 knows that tags 1 and 5 are assigned to channel 0 and tag 7 is assigned to channel 4. Thus, the reader 100 expects a collision on channel 0, but there may also be an unknown tag (in this case, tag 4) that occupies channel 0. Channel 0 indicates the presence of two known tags and the possibility of the presence of one or more unknown tags. Reader 100 did not expect a collision on channel 1 (because no known tags were expecting to select that channel). Here, a collision indicates that there are at least two more unknown tags, possibly more tags. A collision on channel 4 where only tag 7 was expected indicates that there is at least one other unknown tag. Thus, in path # 2, there are four already known tags and at least three (definitely) unknown tags. This is less than the set defined by the first pass, which was at least four known tags and at least four unknown tags, so the reader 100 collects new information in the second pass I did not.

  In path # 3, tag 2 is identified on channel 0. Tag 1 is the only tag that is expected to enter channel 3, so the collision indicates that there is at least one unknown tag there. Tag 7 is the only tag that is expected to enter channel 4, so the collision is there of at least two unknown tags (an unknown tag on channel 3 and an unknown tag on channel 4) Indicates that there is. Tag 3 is also single in this case. Tag 5 was the only tag expected to enter channel 7. A collision indicates that there are at least three unknown tags (counting unknown tags on channels 3, 4, and 7). These also indicate that there are at least 8 tags, as well as the current 5 known tags.

  Path # 4 identifies new tags 4 and 8. Tags 3, 5, and 7 appear in channels that do not collide. Tags 1 and 2 were expected to collide on channel 6, but there could be additional tags. This leaves 7 known tags and at least one unknown tag from the previous experiment.

  There is no new tag identified by path # 5. A collision on channel 5 is not expected, again indicating that there are seven known tags and at least one unknown tag. The same interpretation can be performed from the path # 6 and the path # 7.

In path # 8, tag 6 is identified. All other collisions were expected. Thus, eight tags have been identified and a minimum number is expected from the previous pass. However, there can still be tags hidden in the collision. For example, there may be a tag that has selected channels 1, 0, 4, 6, 3, 1, 1, 5, which will be hidden by other collisions. The probability that the tag has this specific ID is ^1/8 or 6 × 10 ⁻⁸ .

There can also be a tag that selects a channel, eg 2, 4, 4, 6, 5, 4, 5, 6, again with a probability of 6 × 10 ⁻⁸ . In summary, 2 collisions in path # 1, 3 collisions in path # 2, 3 collisions in path # 3, 1 collision in path # 4, 2 collisions in path # 5, 2 collisions in path # 6 , There are 3 collisions in path # 7 and 3 collisions in path # 8, there are 2 × 3 × 3 × 1 × 2 × 2 × 3 × 3 = 648 possible IDs, each with a probability of 6 The probability of a single hidden tag added at x10 ⁻⁸ is 648/8 ⁸ = 38.6 × 10 ⁻⁶ (38.6 ppm). The probability of the two hidden tags being added is even smaller, 64
8 · 647/8 ¹⁶ = 1.5 × 10 ⁻⁹ . In other embodiments, the data is scrambled and, for example, a hidden tag is associated with all or part of other expected products if all other products are groceries. It would be possible to further improve the level of inventory reliability.

The probability of hidden tags can be reduced by continuing the experiment after identifying the minimum number (eight in this case) of expected tags based on collision information. By counting the number of collisions per path and knowing the probability of hidden tags based on the number of channels per path, the reader 100 can meet some confidence level or be a unique channel. It is possible to continue the path until the pattern is exhausted (ID is exhausted). Assuming 648 ^1/8 = 2.246 collisions per pass after two additional passes (total 10 passes), the probability of a single hidden tag is 3.04 × 10 ^{− 6} is possible. After two additional passes (total 12 passes), the probability of a single hidden tag is lowered to 240 × 10 ⁻⁹ . Each time a path is added, the probability of a single hidden tag is reduced by a geometric series approximately 648 ^1/8 /8=0.281 times.

  A flow diagram showing the steps involved in the method without interference cancellation described above is shown in FIG. At start 2210, the system is initialized as having no positive ID or unknown product, which together corresponds to a total of 0 products. After analysis 2230 of the first pass 2220, positive IDs (eg, products 1, 3, 5, and 7 in pass # 1) are recorded and added 2240 to the list of positive IDs. The number of collisions in pass 2250 is also recorded (eg, two collisions in pass # 1). If a collision was expected 2260, there are potentially unknown products that can be revealed in future passes, but there is no sure unknown product. If no collision is expected 2270, two unknown items are added to the unknown list. Thus, the total number of merchandise is estimated 2280 as the number of merchandise identified as positive and the minimum number of unknown merchandise that may have caused the recorded conflict. Assuming that the positive ID is not equal to the estimated total number of items, the total of unknown items is reset to 2295 and another pass 2220 begins. Finally, if the number of positive IDs is equal to the maximum number of already identified IDs plus the number of unknown items, the loop is exited 2290 and a predetermined confidence level 2296 is met.

  So far, no assumptions have been made regarding the temporal variations in channel and received signal level. The “method without collision mitigation” can be applied whether the channel is static or dynamic. In the case of a static channel state, if the power level and phase of the return signal are consistent, more information is available in the reader 100 in the form of the received signal level. So, assuming that the signal level is known in addition to the known channel that the known tag will select in future paths, the tag hidden further in the expected collision It is possible to determine whether or not exists. For example, a collision on channel 0 of path # 2 included two known tags and one unknown tag. If the signal level of the known tag is also known, the total signal level of the collision can be compared with the individual signal level to determine whether additional unknown tags are hidden in the collision. Will. In such an environment, if all of the collisions are accounted for and it is certain that there are no hidden tags, after all tags are identified independently (in this case, 8 passes), the reader 100 will query. It is possible to end.

  Therefore, if the signal level of the identified tag is known, the reliability of inventory accounting is increased. However, with signal level information, it is possible to expect an improvement in collection time rather than simply aborting the query after all known tags have appeared individually. This will be explained in the next section.

Forward Collision Mitigation Example When tags are individually identified, channel selection for all subsequent paths is known at the reader 100 side. If the signal level and phase of the tag are further determined, the contribution of the tag related to the collision can be reduced to zero. From subsequent collisions, the signal from the tag can be essentially removed, and thus can also be actually removed from the population. Consider the experiment shown in FIG. Tags 1, 3, 5, and 7 are identified as positive in path # 1. Assume that the signal level and phase have also been determined.

  On path # 2, tags 1 and 5 are known to transmit data on channel 0. At that known signal level, the tag can be subtracted, leaving only the tag 4 that can currently be identified. Similarly, tag 7 is expected to transmit data over channel 4 on path # 2, and disabling this tag leaves tag 6 intact and identified. Channel 1 still has unresolved collisions and requires at least one other path through this algorithm.

  In path # 3, tag 2 appears alone and is identified. Tag 1 is expected to transmit data on channel 3 and is therefore subtracted, leaving only tag 8 behind, where it can be identified. All other collisions include only known tags, and tag accounting can involve more than eight paths (with the required confidence level) without collision mitigation as shown in FIG. Is completed in three passes through the algorithm with 100% confidence.

  In the static channel, the signal strength of the identified tag can be known with high accuracy. Consider the case of an increased PN channel. In this experiment, the tag selects various code phases of the increased PN sequence 8 chips long. This 8-chip long PN sequence is transmitted as true or inverted for each bit of the tag ID depending on the meaning of a specific ID bit. In reader 100, the correlator in the receiver essentially averages the signal level over 8 chips per bit. This is done for all bits in the ID (eg, 128) and averages 8 × 128 = 1024 samples for a signal to noise ratio average gain of 10 log (1024) = 30 dB. For more practical cases where there are more expected tags and more (> 32) available channels, this gain is greater. With 32 channels and 128 bits, a signal to noise ratio gain of 36 dB is obtained.

Bidirectional Collision Mitigation Example If the reader 100 stores waveform samples from the previous pass, the accounting time may be further improved. If the waveform is stored, it is possible to follow the previous path again and treat it as a subsequent path, from which the previous collision can be invalidated. This is because once a tag is identified, not only all subsequent activity is known, but all previous channel selections and signal levels are also known.

  Consider the example shown in FIG. In path # 1, tags 1, 3, 5, and 7 are identified by both bit pattern and signal level and phase. As with forward collision mitigation, tag 4 can be identified by path # 2, but the effects of tags 1 and 5 can be removed from collisions on channel 0. Similarly, tag 6 can be identified by removing the effect of tag 7 from a collision on channel 4. After pass # 2 and after applying forward collision mitigation, tags 1, 3, 4, 5, 6, and 7 are found.

It is possible to return to the result of pass # 1 after applying forward collision mitigation without using the third pass. If tag 4 is identified in pass # 2, it is possible to remove tag 8 from channel 2 resulting from the first pass and resolve tag 8. If tag 6 is identified in pass # 2, it is possible to remove tag 2 from channel 1 resulting from the first pass and resolve tag 2. In this case, only two passes are required to successfully identify all eight tags. The benefits of forward and bidirectional collision mitigation increase further as the number of channels and tags involved increases.

  Therefore, a one-way communication system using a multi-path transmission algorithm (preferably employing a spread spectrum technique) that exhibits superior performance (eg, read time and capacity) has been described in detail. Incorporating collision mitigation techniques, dynamic channel profiles, and power-on ranges further improves system performance. The described communication system has numerous applications that are not limited to the preferred embodiments and practical examples detailed herein. In addition, the present invention has applications for two-way communication devices, actively powered user devices, and network connection devices without departing from their essential characteristics (as set forth in the claims).

  The present invention may be implemented in other specific forms without departing from the spirit or essential characteristics of the invention. The described embodiments are to be considered in all respects only as illustrative and not as restrictive. The scope of the invention is, therefore, indicated by the appended claims rather than by the foregoing description. All changes that come within the meaning and range of equivalency of the claims are to be embraced within their scope.

FIG. 3 is a high level diagram illustrating multiple source devices communicating with a single destination device according to the present invention. The figure explaining the change method of the data currently recorded on the tag, and the usage method for discriminating a transmission channel, while the operation | movement by this invention is performed. FIG. 4 is a high level diagram illustrating the steps used to scramble data recorded on a tag according to the present invention. FIG. 2 is a high level system diagram illustrating descrambling (scramble analysis method) performed by a plurality of tag communications and readers according to the present invention. FIG. 6 illustrates an iterative algorithm used for tag data scrambling and recovery according to the present invention. The figure which shows the detailed example of the scramble procedure by this invention. FIG. 3 is a high level block diagram of a tag according to the present invention. FIG. 3 is a general flowchart showing an outline of a tag transmission state according to the present invention. 4 is a detailed flowchart showing an outline of a tag transmission state according to the present invention. FIG. 6 illustrates an application that uses electrostatic coupling between a reader and various tags in an exemplary embodiment according to the present invention. The figure which shows the method of producing | generating the channel of the tag communicated based on the data recorded on a tag by this invention. FIG. 4 is a functional block diagram of a simplified tag circuit illustrating path dependency and modulation methods according to the present invention. 1 is a detailed block diagram of a reading device according to the present invention. The figure which shows the example of the high-speed conversion system of the Walsh encoding signal by this invention. The figure which shows the detailed example of the reader-receiver signal process for calculating | requiring the high-speed correlation of the pseudo-noise sequence (pseudonoise sequence) by this invention. FIG. 3 is a simplified functional block diagram of reader signal processing according to the present invention. The figure which shows the example of a waveform at the time of the collision presence by this invention. The figure which shows the some waveform example when a collision does not exist by this invention. 2 is a general flowchart of the operation of the reader according to the present invention. 4 is a detailed flowchart of reader signal processing using forward collision mitigation techniques according to the present invention. The figure which shows the example of inventory accounting when the collision mitigation method is not applied by this invention. 6 is an example of a flow chart of an inventory algorithm when a collision mitigation technique is not applied according to the present invention. The figure which shows the example of the inventory accounting in which the collision mitigation method by this invention is used. The figure which shows the example of the inventory accounting in which the bidirectional | two-way collision mitigation method by this invention is used.

Claims (41)

A receiver for receiving elements of a signal sequence;
A state generator for generating an address sequence and converting between a pseudo-noise sequence and a Walsh sequence;
A device comprising: a storage medium coupled to the receiver and the state generator for storing each element of the signal sequence at an address given according to the address sequence. The device of claim 1, wherein the state generator comprises a linear feedback shift register. The device of claim 2, wherein the linear feedback shift register is a Fibonacci sequence generator. The device of claim 1, wherein the state generator is a second storage medium. The device of claim 1, wherein the pseudo-noise sequence is a specially augmented m-sequence. The device of claim 1, further comprising a processor coupled to the storage medium for performing a transformation on at least a portion of the elements of the signal sequence stored on the storage medium. The device of claim 6, wherein the transform is selected from the group consisting of a fast Hadamard transform (FHT), a fast Walsh transform, and a fast Walsh-Hadamard transform. The device of claim 1, wherein the receiver comprises an analog to digital converter. A collision mitigation method used in a communication system, comprising:
Estimating a signal received on the first channel;
Determining a set of channels used to receive the signal;
Removing the signals from the plurality of signals received on the second channel based on the steps of estimation and determination. The method of claim 9, wherein the estimating step comprises estimating a received signal strength of the signal. The method of claim 9, wherein the step of estimating uses error correction coding. 10. The method of claim 9 , wherein the process of claim 9 is performed iteratively until all signals are determined. The method of claim 9, wherein the signal represents at least a portion of predetermined data stored on a device. The method of claim 9, wherein the first channel and the second channel are the same. The method of claim 9, wherein the first channel and the second channel are different. The method of claim 9, wherein the first channel is orthogonal to the second channel. The method of claim 9, wherein the first channel is pseudo-orthogonal to the second channel. A collision mitigation method used in a multi-path communication system comprising:
Estimating a signal received on a first channel in a given path;
Determining a set of channels used to receive the signal on a previous path and to be used to receive the signal on a subsequent path;
Removing the signals from the plurality of signals received on the second channel based on the steps of estimation and determination. The method of claim 18, wherein the signal is removed from at least one of a previous path and a subsequent path. The method of claim 18, further comprising storing all signals received on each channel in each path. 19. The method of claim 18, wherein in each path, a plurality of devices transmit their respective signals to a common device on the selected channel. The method of claim 18, wherein the estimating step uses error correction coding. 19. The method of claim 18, wherein the process of claim 18 is repeatedly performed until all signals are determined. A collision mitigation method used in a communication system, comprising:
Receiving a signal on a first channel;
Determining a set of channels used to receive the signal;
Estimating the total number of signals in the system based on the number of known signals and the number of colliding signals based on the determining step. 25. The method of claim 24, wherein the steps of claim 24 are iteratively performed until the number of known signals is equal to the estimated total number of signals. 25. The method of claim 24, wherein the steps of claim 24 are performed repeatedly until a predetermined confidence level is obtained. Receiving a signal on the channel;
Estimating the variance of the absolute value of the signal;
Determining, based on the step of estimating, that a collision has occurred on the channel when the estimated variance exceeds a predetermined threshold. 28. The method of claim 27, wherein the predetermined threshold is determined from an average of absolute values of the signal. Receiving a carrier signal; and
Continuously monitoring the carrier signal for a first predetermined condition and a second predetermined condition;
Transmitting the data if the first predetermined condition is satisfied; and
Stopping the transmission of the data if the first predetermined condition is no longer met;
And a step of stopping data transmission when a second predetermined condition is satisfied. 30. The method of claim 29, wherein the first predetermined condition is satisfied when a received power level exceeds a threshold. 30. The method of claim 29, wherein the second predetermined condition is satisfied when a received power level exceeds a threshold. 30. The method of claim 29, wherein the first predetermined condition is met when a predetermined synchronization pulse is received. 30. The method of claim 29, wherein the first and second predetermined conditions are random. A receiver for receiving a carrier signal;
A monitor coupled to the receiver for continuous monitoring of the carrier signal;
A storage medium for storing data;
At least one consisting of the receiver, the monitor, and a transmitter coupled to the storage medium for transmitting at least a portion of the data when first and second conditions are met First device. 35. The at least one first device of claim 34, wherein the first and second conditions of the first device are the same as the first and second conditions of a second device. 36. The at least one first device of claim 35, wherein the first and second devices transmit simultaneously. The at least one first device of claim 34, wherein the first and second conditions of the first device are different from the first and second conditions of the second device. 38. At least one first device according to claim 37, wherein the first and second devices transmit simultaneously. The at least one first device of claim 34, wherein at least one of the first and second conditions is randomly assigned. The at least one first device of claim 34, wherein at least one of the first and second conditions is uniformly distributed. The at least one first device of claim 34, wherein the first and second conditions are met when a power level is within a given range.

Images