CN117113804A

CN117113804A - RFID lost tag identification method based on tree splitting

Info

Publication number: CN117113804A
Application number: CN202310827098.7A
Authority: CN
Inventors: 张莉涓; 范明秋; 何其睿; 宋晓勤; 雷磊; 朱晓浪; 喻春妮; 吴志豪; 陈宇枫
Original assignee: Nanjing University of Aeronautics and Astronautics
Current assignee: Nanjing University of Aeronautics and Astronautics
Priority date: 2023-07-06
Filing date: 2023-07-06
Publication date: 2023-11-24

Abstract

The invention discloses a Tree-splitting-based RFID lost tag identification method (TSMTI, tree-splitting-based Missing Tag Identification Method), which solves the problem that the identification efficiency of the existing lost tag identification method cannot meet the real-time requirement. The method adopts an optimal tree splitting strategy, maps the label to be identified into time slots according to a hash function, and splits each collision time slot into B branches, so that the splitting is carried out continuously until no collision time slot is generated any more. And then comparing the time slot reply condition with a vector calculated in advance by a reader to identify the lost tag. The information pre-calculated by the reader side can be better utilized in a tree splitting mode, and the time slot utilization efficiency is improved. In addition, the method also adopts a bit reply strategy, so that a plurality of tags can be confirmed simultaneously, and the time cost of tag reply is reduced. MATLAB simulation results prove that the method has the advantage of improving the time efficiency of lost label identification.

Description

RFID lost tag identification method based on tree splitting

Technical Field

The invention belongs to the field of Internet of things, and particularly relates to a lost tag identification method suitable for a large-scale RFID system.

Background

Radio frequency identification technology (RFID, radio Frequency Identification) is an emerging automated identification technology, which is one of the important technologies in the internet of things wave. The RFID system is mainly composed of three parts: the system comprises a back-end server, a reader and an electronic tag. Compared with the traditional identification mode, such as a two-dimensional code, the RFID has the advantages of non-contact, non-visual reading, strong anti-interference capability, high reliability, large storage medium capacity and the like. Therefore, the method is widely applied to the fields of inventory management, logistics tracking, medical pharmacy and the like, and can be even deployed in environments with strong interference and other severe environments. When the RFID system is used in inventory management, the reader is deployed in a warehouse, connected to a backend server, and the tag stores key information of the item and is attached to the item being managed. The tag identification protocol is one of the most commonly invoked protocols in RFID systems. Through the tag identification protocol, the reader continuously reads tag information in the communication range, and functions of real-time management, theft prevention and the like are realized.

Currently, in the large-scale RFID system scenario, there are two main types of identification protocols: probability type and deterministic type. Probability-based protocols are used to detect if there are missing tags within a certain threshold, and cannot determine which tag is missing. The deterministic protocol can accurately identify the missing tag and return the tag's ID information. The deterministic protocol consumes longer time than the probabilistic protocol, but has the advantage of being able to be used to track lost tags, so it is more widely used in RFID systems. In recent researches, a great deal of missing tag identification protocols are proposed, so that RFID technology is rapidly developed. The THP (two hash protocol) protocol uses an allocation of two hash functions to identify lost tags. To be suitable for use in an actual scenario, part of the method is by building two vectors: the expected vector and the actual vector are compared, the interference of unknown labels is eliminated in the first stage, and the missing labels are identified in the second stage, so that the identification efficiency and the reliability of the system are greatly improved. However, this way of identification by hash mapping and vector contrast, the efficiency improvement is limited.

To further improve efficiency, more and more researchers have applied bit detection techniques to speed up the recognition rate. The application of the bit detection technology realizes the utilization of information in collision time slots, and the reader can collect the reply information of the tag in the collision time slots, so that the time efficiency is further improved. Compared with the traditional method, the method for coordinating and pairing the collision time slots can simultaneously identify two tags in one time slot, and improves the time slot efficiency by more than 50%. The collision resolution-based method further improves the recognition efficiency through collision coordination and bit detection. However, the above-described identification method has the following problems: (1) The communication protocol between the existing reader and the tag often adopts a traditional ALOHA structure, the frame identification efficiency of each protocol is limited to 1/e, and improvement of the efficiency is needed; (2) The reader can judge the label information under the collision time slot in advance, however, the information is not fully utilized, and the utilization of the time slot still has a lifting space.

The invention considers the scene that the known tags exist in a large-scale RFID system and part of the known tags exist and part of the known tags are lost. In order to quickly identify the lost tag, a method capable of identifying the lost tag and fully utilizing the time slot is required to be designed, so that the identification efficiency of a system is improved, and the identification time cost is reduced.

Disclosure of Invention

The purpose of the invention is that: aiming at the scene that a large number of known tags exist in the single reader, wherein part of tags are lost, the RFID lost tag identification method based on tree splitting is provided, and the lost tags are rapidly identified through the splitting process of collision time slots. We describe this problem in the following way: in a system where a known tag set is E and a lost tag set is M and M E, where the existing tag set is p=n-M, the protocol needs to identify all the lost tags in the set M in as short an execution time as possible. In order to achieve the object, the invention adopts the following steps:

step 1: at the reader end, an expected indication vector IV is established by adopting a Hash function mapping mode ₁ . Setting an initial frame length f ₁ Generating random seed R ₁ The reader performs hash function mapping according to the known tag ID information in the background server to obtain an expected state vector IV ₁ . There are three cases according to label mapping: empty slots, single slots, and collision slots, each slot in the corresponding vector has a value of "0", "10", "11". The time slot mapped by each tag changes the state of the corresponding position of the vector to obtain the expected state vector IV ₁ . The reader can then learn the expected mapping of time slots for all known tags.

Step 2: at the reading end, the tags under the collision time slot are mapped to B branches by adopting a tree splitting mode, and an expected vector IV is established ₂ . First, the reader generates a new parameter B and a hash seed R ₂ . And according to the mapping condition of the first round, performing hash mapping again on the tags under the collision time slot. The labels in the same collision time slot are a group, hash mapping is carried out on B time slots to obtain the current sub-indication vector, and the vector is also composed of three numerical values of 0, 10 and 11 according to mapping conditions. All tags under collision time slots are mapped in the same way, and the generated sub-indication vectors are spliced into an indication vector IV ₂ . The tags under the collision slot are then split in this step until no collision slot is generated. Finally, a set of the required random seeds is obtained, and the identification process is carried out.

Step 3: reader broadcast parameter f ₁ ，R ₁ And an indication vector IV ₁ After the label receives the parameters, the label receives the parameters according to the hash functionAnd (3) recovering. If the label is mapped to the time slot, the corresponding indication vector IV ₂ The position of "10" in (C) will generate a length L ₁ Where the x-th bit is a "1" and the remainder are "0". The value of L is the indication vector IV ₁ The total number of "10" in (a), the value of x is the tag map corresponding IV ₁ 10 "before the slot of (c) and 1. The tags remain silent after reply and do not participate in the subsequent identification process. If the position corresponding to 11 is the position corresponding to 11, the counter Ac of the self is set according to the index value of the collision time slot. The tag remains silent in the current frame waiting to participate in the next frame. After receiving the reply, the reader compares the decoded reply with the expected indication vector, identifies the tag and confirms the existence of the tag.

Step 4: reader broadcasts B, R ₂ And an indication vector IV ₂ The labels under the same collision time slot are a group, the index value is calculated and distributed to B branches under the same group, and the reply is carried out according to the mapping condition. If the label is mapped to the time slot, the corresponding indication vector IV ₂ The position of "10" in (3) will generate a length L ₂ Where the x-th bit is a "1" and the remainder are "0". L (L) ₂ The value of which is the indication vector IV ₂ The total number of "10" in (a), the value of x is the tag map corresponding IV ₂ 10 "before the position of (c) plus 1. And the tag remains silent and no longer participates in the subsequent identification process. If the position corresponding to 11 is the position corresponding to 11, the counter Ac of the self is set according to the index value of the collision time slot. The tag remains silent in the current frame waiting to participate in the next frame. After receiving the reply, the reader compares the decoded reply with the expected indication vector, identifies the tag and confirms the existence of the tag.

Step 5: and (3) repeating the operations of the steps 3-4 by the reader and the tags until the reader broadcasts all parameters and all tags are identified and collision time slots are not generated.

The RFID lost label identification method based on tree splitting is realized in MATLAB, and the simulation result verifies the superiority of the method. In the simulation experiment, the ID length of the tag is assumed to be 96 bits, and the communication protocol between the reader and the tag is based on an FSA algorithm. In fig. 1, the RFID system consists of a server, a single reader, and existing tags and missing tags. In fig. 2, the known tag set e= { T1, T2, T3, T4, T5, T6, T7, T8, T9, T10}, the missing tags M E, m= { T2, T5, T8}. Fig. 3 shows the total execution time of the protocol with an environment of n=1000 to 5000, a step size of 200, and a missing tag ratio of 0.2. Fig. 4 shows the comparison result of n=3000, the missing tag rate from the execution time of the 0.1-1 system and other missing tag identification methods. The comparison result illustrates the high efficiency of the RFID lost label identification method based on tree splitting.

Drawings

FIG. 1 is a schematic diagram of an RFID system model of the present invention;

FIG. 2 is a schematic diagram of a tree-splitting-based RFID missing tag identification method of the present invention;

FIG. 3 is a schematic diagram of the time overhead for different known tag numbers of the present invention;

fig. 4 is a schematic diagram of the time overhead at different lost tag ratios of the present invention.

Detailed Description

The invention is described in further detail below with reference to the drawings and examples.

In the following description, the present specification will make the tag identification, abbreviated as TSMTI (Tree splitting missing tag identification), which is applicable to the present invention for a large-scale RFID system. The TSMTI first sets the following system parameters:

1. known tag sets: E. in the RFID system, it is necessary to manage and identify a tag set stored before, and a background server stores information such as 96-bit ID. N is the total number of known tags.

2. Missing tag set: m. The RFID system identifies the lost tag and stores the relevant information before it is identified. m is the total number of unknown tags.

3. Initial frame length: f (f) ₁ The initial frame length used for the identification process is β×n.

4. Frame length influencing factor: beta.

5. Random seed: r is a generated random seed, and the length is 16-bit.

6. Hash function: h (ID) _j R), H () represents a hash function, ID _j ID information indicating the j-th known tag.

7. Tree branch number: and B, determining the number of time slots into which one collision time slot is split.

8. Expected indication vector: IV, a vector calculated in advance in the reader based on the known tag information. The vector length is f bits, and each bit of the initial state is "0". There are three values of "0", "10" and "11", with "0" representing an empty slot, i.e. a slot to which no tag is mapped. "10" means a single slot, i.e. only one label map. "11" indicates a collision slot, with multiple tag mappings.

9. Actual indication vector: AV, the vector built according to the tag actual reply. There are two values of "0" and "1," 0 "indicating an empty slot, i.e., no tag replies. "1" indicates a non-empty slot, i.e., a labeled reply.

Based on the above conditions, the TSMTI provided by the invention realizes a specific scheme in MATLAB, and the realization result proves the effectiveness of the method. The specific implementation steps of TSMTI are as follows:

step 1: at the reader end, an expected indication vector IV is established by adopting a Hash function mapping mode ₁ . The reader establishes an indication vector, and the initial frame length f ₁ Is composed of beta-N time slots (beta is the frame length influencing factor, N is the number of known labels), and generates random seed R ₁ . Subsequently, hash function mapping is performed based on all known tag ID information in the background server, e.g., the t-th tag calculates i=h (ID _t ，R ₁ )mod f ₁ +1. If there is a position to which the tag is mapped, the corresponding indication vector time slot is set to be 0; only one tag maps to a position, and the corresponding indication vector time slot is set to be 10; with a plurality of tags mapped to positions, the corresponding indicator vector slot is set to "11", and the expected state vector IV is finally obtained ₁ . The reader is thus made aware of allThe mapping situation expected for the time slot of the tag is known.

Step 2: at the reading end, the tags under the collision time slot are mapped to B branches by adopting a tree splitting mode, and an expected vector IV is established ₂ . The reader generates a new parameter B and a hash seed R ₂ . And according to the mapping condition of the first round, performing hash mapping again on the tags under the collision time slot. The tags are grouped according to collision time slots, and the tags in the same collision time slot are grouped. Tags of the same group will calculate I=h (ID) from a hash function, e.g. the t-th tag _t ，R ₂ ) mod B+1 maps to B slots to obtain the current sub-indicator vector. According to step 1, the sub-indication vector is composed of three values of "0", "10" and "11" for each slot according to the same mapping condition. All the tags under the collision time slots are mapped identically and are independent of each other, c sub-indication vectors are generated, and c represents the total number of expected collision time slots of the previous frame. The reader concatenates all sub-vectors into an indication vector IV ₂ The total time slot number is f ₂ =c×b. Then, the division is continuously carried out according to the collision time slot until no collision time slot exists, and the random seed R of the ith frame is obtained _i The identification process is entered.

Step 3: the reader broadcasts the parameter f to the tag by commanding Query ₁ ，R ₁ And an indication vector IV ₁ After receiving the parameters, the tag replies according to the hash function. For example, the t-th tag, i=h (ID _t ，R)mod f ₁ +1, if corresponding time slot IV ₁ [i]= "10", a length L will be generated ₁ Where the x-th bit is a "1" and the remainder are "0". The value of L is the indication vector IV ₁ The total number of "10" in (a), the value of x is the tag map corresponding IV ₁ 10 "before the slot of (c) and 1. The tag remains silent after reply and no longer participates in the subsequent identification process. If it corresponds to IV ₁ [i]= "11", then the own counter Ac is set according to the total number of the previous "11" s of the collision slot where it is located plus 1. The tag then remains silent in the current frame waiting to participate in the identification of the next frame. The reader receives the tag replyThen, according to Manchester decoding, an actual reply vector AV is obtained ₁ 。AV ₁ The bit in (a) is "1", and the corresponding indication vector IV ₁ The corresponding single time slot in the existing label set P is added with the label in the corresponding single time slot in the existing label set P, and the corresponding known label is lost and added with the identified lost label set M.

Step 4: the reader broadcasts B, R to the tags by commanding Query ₂ And an indication vector IV ₂ . According to the setting of the above steps, the tags are grouped according to the value of the own counter Ac, and the tags under the same collision time slot are grouped. The index value of the tags in the same group is calculated through a hash function H (ID, R) mod B+1, and the index values are distributed into B time slots in the same group to form sub-indication vectors. The tag then replies according to the mapping. If the corresponding time slot IV ₂ [i]= "10", a length L will be generated ₂ Where the x-th bit is a "1" and the remainder are "0". L (L) ₂ The value of which is the indication vector IV ₂ The total number of "10" in (a), the value of x is the tag map corresponding IV ₂ 10 "before the position of (c) plus 1. The tag remains silent after reply and no longer participates in the subsequent identification process. If it corresponds to IV ₂ [i]= "11", then the own counter Ac is set according to the total number of the previous "11" s of the collision slot where it is located plus 1. The tag remains silent in the current frame waiting to participate in the next frame. After receiving the tag reply, the reader obtains an actual reply vector AV according to Manchester decoding ₂ 。AV ₂ The bit in (a) is "1", and the corresponding indication vector IV ₂ The corresponding single time slot in the existing label set P is added with the label in the corresponding single time slot in the existing label set P, and the corresponding known label is lost and added with the identified lost label set M.

Step 5: and (3) repeating the operations of the steps 3-4 by the reader and the tags until the reader broadcasts all parameters, and after the tags are recovered, all the tags are identified and collision time slots are not generated any more.

What is not described in detail in the present application belongs to the prior art known to those skilled in the art.

Claims

1. The RFID missing tag identification method based on tree splitting comprises the following steps:

Step 3: reader broadcast parameter f ₁ ，R ₁ And an indication vector IV ₁ After receiving the parameters, the tag replies according to the hash function. If the label is mapped to the time slot, the corresponding indication vector IV ₂ The position of "10" in (C) will generate a length L ₁ Wherein the x-th bit is "1" and the remainder are"0". The value of L is the indication vector IV ₁ The total number of "10" in (a), the value of x is the tag map corresponding IV ₁ 10 "before the slot of (c) and 1. The tags remain silent after reply and do not participate in the subsequent identification process. If the position corresponding to 11 is the position corresponding to 11, the counter Ac of the self is set according to the index value of the collision time slot. The tag remains silent in the current frame waiting to participate in the next frame. After receiving the reply, the reader compares the decoded reply with the expected indication vector, identifies the tag and confirms the existence of the tag.

And (3) repeating the operations of the steps 3-4 by the reader and the tags until the reader broadcasts all parameters, and after the tags are recovered, all the tags are identified and collision time slots are not generated any more.

2. The method according to claim 1, characterized in that the specific method for identifying lost tags based on tree splitting is:

at the reading end, the tags under the collision time slot are mapped to B branches by adopting a tree splitting mode, and an expected vector IV is established ₂ . The reader generates new parametersB and hash seed R ₂ . And according to the mapping condition of the first round, performing hash mapping again on the tags under the collision time slot. The tags are grouped according to collision time slots, and the tags in the same collision time slot are grouped. Tags of the same group will calculate I=h (ID) from a hash function, e.g. the t-th tag _t ，R ₂ ) mod B+1 maps to B slots to obtain the current sub-indicator vector. According to step 1, the sub-indication vector is composed of three values of "0", "10" and "11" for each slot according to the same mapping condition. All the tags under the collision time slots are mapped identically and are independent of each other, c sub-indication vectors are generated, and c represents the total number of expected collision time slots of the previous frame. The reader concatenates all sub-vectors into an indication vector IV ₂ The total time slot number is f ₂ =c×b. Then, the division is continuously carried out according to the collision time slot until no collision time slot exists, and the random seed R of the ith frame is obtained _i The identification process is entered.

The reader broadcasts the parameter f to the tag by commanding Query ₁ ，R ₁ And an indication vector IV ₁ After receiving the parameters, the tag replies according to the hash function. For example, the t-th tag, i=h (ID _t ，R)modf ₁ +1, if corresponding time slot IV ₁ [i]= "10", a length L will be generated ₁ Where the x-th bit is a "1" and the remainder are "0". The value of L is the indication vector IV ₁ The total number of "10" in (a), the value of x is the tag map corresponding IV ₁ 10 "before the slot of (c) and 1. The tag remains silent after reply and no longer participates in the subsequent identification process. If it corresponds to IV ₁ [i]= "11", then the own counter Ac is set according to the total number of the previous "11" s of the collision slot where it is located plus 1. The tag then remains silent in the current frame waiting to participate in the identification of the next frame. After receiving the tag reply, the reader obtains an actual reply vector AV according to Manchester decoding ₁ 。AV ₁ The bit in (a) is "1", and the corresponding indication vector IV ₁ The presence of a tag in a corresponding single slot in, adding to the existing set of tags P, a position of "0",the corresponding known tag is lost and added to the identified lost tag set M.

The reader broadcasts B, R to the tags by commanding Query ₂ And an indication vector IV ₂ . According to the setting of the above steps, the tags are grouped according to the value of the own counter Ac, and the tags under the same collision time slot are grouped. The index value of the tags in the same group is calculated through a hash function H (ID, R) mod B+1, and the index values are distributed into B time slots in the same group to form sub-indication vectors. The tag then replies according to the mapping. If the corresponding time slot IV ₂ [i]= "10", a length L will be generated ₂ Where the x-th bit is a "1" and the remainder are "0". L (L) ₂ The value of which is the indication vector IV ₂ The total number of "10" in (a), the value of x is the tag map corresponding IV ₂ 10 "before the position of (c) plus 1. The tag remains silent after reply and no longer participates in the subsequent identification process. If it corresponds to IV ₂ [i]= "11", then the own counter Ac is set according to the total number of the previous "11" s of the collision slot where it is located plus 1. The tag remains silent in the current frame waiting to participate in the next frame. After receiving the tag reply, the reader obtains an actual reply vector AV according to Manchester decoding ₂ 。AV ₂ The bit in (a) is "1", and the corresponding indication vector IV ₂ The corresponding single time slot in the existing label set P is added with the label in the corresponding single time slot in the existing label set P, and the corresponding known label is lost and added with the identified lost label set M.