Improved mobile radio frequency identification tag anti-collision method based on time slot scheduling technology

Abstract

The invention provides an improved RFID tag anti-collision method based on a time slot scheduling technology, and belongs to the field of mobile radio frequency identification (RFID) tag anti-collision algorithms. According to the method, the thought of first arrival and first service is introduced, so that a tag identification algorithm in a mobile RFID scene is improved, and the tag loss rate of a mobile RFID system can be effectively reduced. On the basis of a traditional label estimation algorithm, a label estimation algorithm based on a KAN network is provided, a prediction model is trained based on time sequence data obtained through a dynamic frame time slot Aloha algorithm, and estimation of the number of labels is achieved. Meanwhile, the algorithm introduces the thought of first arrival and first service, defines the concept of priority label coverage rate (PTCR), classifies the labels, fully identifies the preferentially identified labels by adopting a collision tree algorithm, and adopts a mobile frame time slot Aloha protocol for common labels, and the simulation result shows that the algorithm can effectively improve the throughput rate of the RFID system, and has the advantages of being high in reliability and high in reliability. And the loss rate of the label and the communication consumption of the system are reduced.

Description

Technical Field

[0001] The present invention relates to the field of radio frequency identification (RFID) technology, and in particular to an improved mobile radio frequency identification tag anti-collision method based on time slot scheduling technology. Background Art

[0002] Radio Frequency Identification (RFID) technology is widely used in the current era of the Internet of Things. Compared with traditional barcode technology, RFID has obvious advantages in convenience, stability, and security. Therefore, this technology is widely used in transportation, medical care, industry, logistics and other fields.

[0003] In the RFID system, the communication between the reader and the tag is achieved through wireless signal transmission. After the tag enters the working range of the reader, it will send its own data encoding to the reader. However, when there are multiple tags within the reader's recognition range, these tags will respond to the reader at the same time, making it impossible for the reader to correctly receive the valid information of the tags, thus greatly reducing the communication efficiency of the RFID system. This phenomenon is called tag collision. With the widespread application and development of RFID technology, the research on mobile RFID systems has become very important.

[0004] In the mobile RFID scenario, new tags constantly enter and leave the interactive environment, and tags only stay within the working range of the reader for a certain period of time. The traditional mobile frame time slot algorithm based on Aloha does not take into account the entry and exit of tags, and the frame length estimation algorithm is relatively complex and changeable. For the deterministic algorithm based on the tree structure, the entry of tags will destroy its tree structure, resulting in a significant increase in algorithm complexity and low efficiency. Therefore, it is necessary to develop a new mobile radio frequency identification tag anti-collision method, which can effectively reduce the tag loss rate of the mobile RFID system, improve the system throughput and recognition efficiency, and at the same time, the use of the KAN network can generalize the frame length estimation function and improve the accuracy of frame length prediction, so as to adapt to a wider range of scenario requirements and improve the performance and reliability of the system. Summary of the invention

[0005] The purpose of the present invention is to solve the problem of low efficiency of existing tag recognition methods in processing tag mobile environments, and to provide an improved mobile radio frequency identification tag anti-collision method based on time slot scheduling technology, which can effectively deal with the problems of tag loss and low recognition rate in a single reader multi-mobile tag scenario.

[0006] To achieve the above object, the present invention proposes an improved mobile radio frequency identification tag anti-collision method based on time slot scheduling technology. First, the idea of ​​first-come-first-served is introduced, the concept of priority tag coverage rate (PTCR) is defined, the tags are classified, the collision tree algorithm is used to fully identify the priority tags, the mobile frame time slot Aloha protocol is used for ordinary tags, and finally the calculation and prediction process of the frame length is optimized using the KAN network.

[0007] The technical solution proposed by the present invention includes the following parts:

[0008] 1Time slot allocation:

[0009] 1.1 Based on the characteristics of the mobile RFID system that urgently needs to reduce the tag loss rate, the present invention defines the concepts of priority tags and priority tag coverage to ensure that tags that are about to leave the reader's working range can be identified first to reduce the phenomenon of system tag loss.

[0010] 1.2 The present invention follows the idea of ​​first-come, first-served to determine the priority of the tag. First-come, first-served is a common service allocation principle, which is widely used in many fields, such as queuing systems, task scheduling, etc. Under this principle, the first request or task to arrive will be processed first, and the subsequent requests will be processed in the order of arrival. This principle is simple and easy to implement.

[0011] 2 Label classification and recognition:

[0012] 2.1 Based on the definition of priority tags and the characteristics of mobile RFID scenarios, the present invention classifies tags into priority tags that require priority identification and ordinary tags that do not require priority identification.

[0013] 2.2 For different types of tags, the present invention adopts different types of algorithms for classification and identification, which ensures the tag loss rate of the system and also improves the throughput rate of the system.

[0014] 3KAN network generalization frame length prediction:

[0015] Kolmogorov-Arnold Networks (KAN for short) is a new type of neural network architecture, which is built on the Kolmogorov-Arnold representation theorem and provides a deep learning model that is different from the traditional multi-layer perceptron (MLP). The core feature of KAN is that it transfers the activation function from the nodes of the neural network to the edges, and uses learnable functions (such as B-splines) to replace fixed activation functions, thereby eliminating the dependence on linear weight matrices.

[0016] In the dynamic frame-slotted Aloha anti-collision algorithm, the frame length prediction functions are complex and diverse, and there are significant differences in accuracy. The present invention introduces a KAN network model to optimize the prediction of the frame length based on the data obtained in the past, and at the same time improves the generalization ability of the calculation process.

[0017] The present invention introduces the idea of first-come, first-served, defines the concept of priority tag coverage rate (PTCR), classifies the tags, fully identifies the tags with priority recognition using the collision tree algorithm, and uses the moving frame-slotted Aloha protocol for ordinary tags, which can effectively reduce the tag loss rate of the mobile RFID system, improve the throughput rate and recognition efficiency of the system. At the same time, the present invention uses the KAN network to generalize the frame length prediction function and improve the accuracy of frame length prediction, so as to adapt to the needs of a wider range of scenarios and improve the performance and reliability of the system. Brief Description of the Drawings

[0018] Figure 1 is the overall flowchart of the method of an embodiment of the present disclosure;

[0019] Figure 2 is the comparative simulation result according to an embodiment of the present disclosure. Detailed Embodiments

[0020] The present invention provides an improved mobile radio frequency identification tag anti-collision method based on time slot scheduling technology, introduces the idea of first-come, first-served, defines the concept of priority tag coverage rate (PTCR), classifies the tags, fully identifies the tags with priority recognition using the collision tree algorithm, and uses the moving frame-slotted Aloha protocol for ordinary tags, which can effectively reduce the tag loss rate of the mobile RFID system, improve the throughput rate and recognition efficiency of the system. At the same time, the present invention uses the KAN network to generalize the frame length prediction function and improve the accuracy of frame length prediction.

[0021] Step 1, define the priority tag and the priority tag coverage rate: Define the priority tag: that is, the number of stay time slots of the tag in the reader exceeds the threshold set by the mobile RFID system, then the tag in this state is called a priority tag, and the mathematical expression is: TagSlotCounter / MaxStaySlot < Limit; Define the priority tag coverage rate: that is, the ratio of the number of time slots occupied by the priority tag to the total number of time slots in the frame, and this ratio is called the priority tag coverage rate, and the mathematical expression is: PriorTagRate = PriorSlotCount / Frame

[0022] Step 2, identify the priority tags, and the identification process includes the following situations:

[0023] In case 2.1, if there is only one tag that needs to be identified first, the tag directly communicates with the RFID reader and sends its own ID information, which means that the tag is successfully identified;

[0024] Case 2.2: If there is no tag that needs to be identified first, it means that after the last frame, the stay status of all tags is relatively optimistic, and there is no need for priority identification. Ordinary random identification can be used;

[0025] In case 2.3, if there are multiple tags that need to be identified first, appropriate time slots are allocated according to the priority tag coverage, and a deterministic tag anti-collision algorithm, that is, a collision tree algorithm, is used to identify these tags in a timely manner.

[0026] The method according to claim 1 is characterized in that the process of randomly selecting a time slot by a tag is as follows:

[0027] The unrecognized tag in the current frame randomly selects a time slot according to the frame length of the current frame as the time slot for the tag to send its own ID. The mathematical expression is: slotID = random (0, Frame)

[0028] The method according to claim 1, characterized in that the frame cycle identification step comprises:

[0029] Step 5.1, judging the time slot status: judging the current time slot status according to whether there is a tag response in each time slot and whether there is a signal conflict.

[0030] Step 5.2, determine whether new tags are added: According to the inflow speed of tags and the current time slot time, determine whether new tags are added. If so, put them into the tag sequence, but the time slot number is only assigned at the beginning of the next frame.

[0031] Step 5.3, determine whether any unrecognized tag has left: determine whether any unrecognized tag has left based on the time slot counter of the unrecognized tag, and if so, remove it from the tag sequence.

[0032] Step 5.4, update the value of the tag time slot counter: add 1 to the value of the unidentified tag time slot counter, and determine whether the tag is a priority tag according to the concept of defined priority tags. If so, the number of time slots occupied by the priority tag is also increased by 1.

[0033] The method according to claim 1 is characterized in that the steps of predicting the number of labels and calculating the priority label coverage are:

[0034] Step 6.1, tag number prediction: Based on the statistics of the number of successful time slots, collision time slots and empty time slots in the current frame, as well as the tag addition rate and the maximum tag residence time, the KAN network is used to predict the number of unrecognized tags. The mathematical expression is: ExpectFrame = f(Success, Collision, Empty, Increase, MaxSlot, OldFrame).

[0035] Step 6.2, calculate the priority tag coverage rate: According to the priority tag coverage rate of the current frame obtained by statistics, the mathematical expression is: PriorTagRate = PriorSlotCount / Frame.

[0036] experiment

[0037] In order to verify the effectiveness of the method proposed in this invention, simulation experiments were conducted on three key indicators: system tag loss rate, throughput rate and system communication consumption. The initial number of tags and tag IDs were the same in each round of experiments. By changing the tag inflow speed, the method was compared with the traditional algorithms Vogt_DFSA and DFSA-CN. In these three key indicators, this method is better than the traditional algorithms Vogt_DFSA and DFSA-CN, which do not guarantee that tags that are about to leave the reader's working range can be identified first.

Claims

1. An improved mobile radio frequency identification tag anti-collision method based on time slot scheduling technology, characterized in that: The steps include: Step 1: Define the concepts of priority tag, priority tag coverage and priority time slot, and count the priority identification tags in the current frame. Step 2: According to the concept of priority labels defined above, determine the number of priority labels in the current system and classify and identify them according to the situation. Step 3: After the above special tag identification is completed, the reader broadcasts the command to all unidentified tags. Each tag that receives the command is assigned an indefinite time slot number and waits for the target time slot to arrive. Step 4: The frame cycle starts. The system polls and broadcasts the time slot information. The tag of the time slot is selected and its ID and other information are sent to the reader. The reader determines and counts the status of the time slot based on the received information and broadcasts a command to modify the time slot counter information of the tag. Step 5: After the frame cycle ends, the reader calculates the priority tag coverage of the current frame based on the statistically obtained number of successful time slots and collision time slots, uses the KAN network model to predict the number of unrecognized tags, and repeats the above steps.

2. The method according to claim 1, characterized in that The steps to define priority labels and priority label coverage include: Step 2.1, that is, the number of time slots that the tag stays in the reader, TagSlotCounter, exceeds the threshold Limit set by the mobile RFID system, then the tag in this state becomes a priority tag. Such a tag meets TagSlotCounter / MaxStaySlot≥Limit Where MaxStaySlot is the maximum number of time slots that a tag can stay. Assuming that the speed of adding tags is S, in a mobile RFID system, tags will only stay in fixed time slots. The identification process is constantly accompanied by the addition and departure of tags, and the number of tags is also constantly changing. Therefore, this paper assumes that the reader has scanned k-1 rounds. After the first k-1 rounds of identification, the estimated number of tags is n, the frame length is f, and the number of priority tags in the kth round is p. k , the tag enters the working range of the reader at time slot t0 and starts sending signals at round k0. It can be seen that for p k The number of time slots s required for a tag to perform priority identification k =2p k -1, when the number of tags is equal to the frame length, the DFSA algorithm throughput reaches its maximum value. Therefore, at the beginning of the kth round of recognition, the relationship between the actual number of tags and the predicted number of tags is At the beginning of round k, the number of time slots that the tag stays at the reader Therefore, if we want to analyze the situation of signing out, k satisfies S k ≥m Where m is the maximum time slot that the tag can stay within the system working range. If the tag is a priority tag, the number of time slots it needs to stay within the reader's operating range s=L*m Where L represents the threshold set by the RFID system, Step 2.2, define the priority tag coverage rate: that is, the ratio of the number of time slots occupied by the priority tag in the current round of ordinary identification (PriorSlotCount) to the total number of time slots in the frame (Frame). This ratio is called the Priority Tag Count Rate (PTCR), and the mathematical expression is: PTCR=PriorSlotCount / Frame Step 2.3, define the priority time slot: The time slot selected by the priority tag is called the priority time slot. This definition is limited to the current frame polling. The priority time slot may change in each round of identification. It can be seen that the probability of a tag being identified in a certain time slot during the ordinary random identification process is In the normal identification process, if the selected time slot has at least one other tag that selects the same time slot, the tag cannot be identified in this polling, and the probability is It can be seen that the probability of a label becoming a priority label is:

3. The method according to claim 1, characterized in that The recognition of priority labels includes the following situations: Case 3.1: If the number of tags that respond to the signal does not exceed the threshold, it means that after the previous frame, the stay of all tags is relatively optimistic and no priority recognition is required. The current frame only needs to use ordinary random recognition; In case 3.2, if there are multiple tags that need to be identified first, appropriate time slots are allocated according to the priority tag coverage ratio PTCR, and a deterministic tree-based anti-collision algorithm is used, that is, a deterministic algorithm is used to identify these tags in a timely manner.

4. The method according to claim 1, characterized in that: The frame cycle identification step comprises: Step 4.1, determine the time slot status: determine the current time slot status based on whether there is a tag response in each time slot and whether there is a signal conflict. Step 4.2, determine whether there are new tags added: According to the inflow speed of the tag and the current time slot time, determine whether there are new tags added. If so, put it into the tag sequence, but only assign the time slot number at the beginning of the next frame. From the recognition process of the algorithm, it can be seen that the total number of new tags added after the kth round of recognition is Step 4.3, determine whether there are unrecognized tags leaving: According to the time slot counter of the unrecognized tag, determine whether there are unrecognized tags leaving. If so, remove them from the tag sequence. The set of tags that leave in the kth round of priority recognition is called Tprior(k). Such tags satisfy S k-1 ≤s S k-1 +2p k -1≥m The set of tags that leave during the kth round of normal recognition is called Tnormal(k). Such tags satisfy S k-1 ≤s S k-1 +2p k -1+f k -t1≥m Where t1 represents the time slot number selected by the tag in the kth round of ordinary recognition Step 4.4, update the value of the tag time slot counter: add 1 to the value of the unidentified tag time slot counter, and determine whether the tag is a priority tag based on the concept of defined priority tags. If so, the number of time slots occupied by the priority tag is also increased by 1.

5. The method according to claim 1, characterized in that The steps of predicting the number of labels and calculating the priority label coverage are as follows: Step 5.1, tag number prediction: Based on the statistics of the number of successful time slots, collision time slots and empty time slots in the current frame, as well as the tag addition speed and the maximum tag residence time, the KAN network model is used to predict the number of unrecognized tags in the system. The core formula of the KAN network is: in is a univariate continuous function defined on the interval [0, 1], It is a function defined on the real number domain. In the KAN architecture, each connection is defined by a single variable function, which directly maps the input to the output without the need for intermediate weighted sums and subsequent general activations. This design provides a more flexible and customized method for data conversion, making KAN robust when processing high-dimensional data and able to form smooth multidimensional surfaces. In addition, KAN also introduces the MultKAN model, which adds a multiplication layer to the original KAN to more clearly reveal the multiplication structure in the data and enhance the interpretability and expressiveness of the model. KAN uses this theory to replace the traditional fixed activation function with a learnable function (such as B-spline) at the edge of the network, thereby improving the accuracy and interpretability of the model. Step 5.2, calculate the priority label coverage, use the KAN network model to predict the number of remaining labels, and determine whether priority identification is required in the next round based on the obtained coverage.

Images