High-density RFID tag group scanning system based on multi-antenna array and beamforming

Abstract

The present invention provides a high-density RFID tag group scanning system based on a multi-antenna array and beamforming, and relates to the field of Internet of Things technology. The system includes a multi-antenna array, an RFID reader / writer, a beamforming unit, a signal processing unit, a feedback unit, and a data management unit. Through digital beamforming technology, the system dynamically adjusts the direction and width of the beam to achieve accurate scanning of the target area; combined with real-time signal feedback, the density peak clustering algorithm and the gradient descent algorithm are used to optimize the beam parameters to improve recognition efficiency. The system further uses an anti-collision algorithm to dynamically adjust frame parameters, and performs logical grouping based on the spatial position and type characteristics of the tag, significantly reducing signal collision and missed reading rates. At the same time, the data management unit classifies, stores and outputs the tag data, providing intelligent management capabilities for multi-scenario applications.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to a high-density RFID tag group scanning system based on multi-antenna array and beam forming. Background Art

[0002] With the rapid development of Internet of Things technology, RFID technology has been widely used in logistics, warehousing, production line management and other fields. In a high-density tag environment, efficient and accurate reading and management of a large number of RFID tags has become a key technical requirement. Traditional RFID tag recognition methods mostly rely on full coverage scanning, which usually takes a lot of time and computing resources. In scenarios where tags are densely distributed, signal conflicts frequently occur, resulting in a significant decrease in tag recognition rate.

[0003] For this purpose, for example, the Chinese patent application with publication number CN117787318A discloses an RFID tag group pairing method and system, which uses RFID group scanning technology to electronically manage port resources by installing trigger tags and detection tags on the ports and optical fiber connectors of network devices, and realizes the identification and inspection management of port status through pairing relationships. This method can quickly locate ports and fault locations, reduce manual registration costs, and improve resource management efficiency.

[0004] However, the above methods have the following shortcomings: they are mainly used in specific scenarios of port resource management, and their functions are relatively simple, making it difficult to meet the needs of fast identification and complex data processing in multi-tag high-density environments. In addition, the identification of tags in this method only relies on simple changes in the tag state, and lacks further optimization of tag distribution and signal characteristics in complex scenarios, which may lead to inefficiency or missed reading in large-scale tag reading scenarios. Summary of the invention

[0005] In view of the shortcomings of the prior art, the present invention provides a high-density RFID tag group scanning system based on a multi-antenna array and beamforming, the system comprising: a multi-antenna array for receiving a response signal from an RFID tag;

[0006] An RFID reader / writer, connected to the multi-antenna array, configured to transmit an inquiry signal to the RFID tag and receive a response signal transmitted by the multi-antenna array;

[0007] A beamforming unit connected to the RFID reader and the multi-antenna array, controlling the amplitude and phase distribution of the multi-antenna array, generating a directional beam, and scanning a target area containing an RFID tag;

[0008] A signal processing unit connected to the RFID reader / writer receives and parses response signals of multiple RFID tags, parses and distinguishes the response signals of multiple RFID tags using an anti-collision algorithm, and generates tag data and real-time signal parameters, wherein the signal processing unit encodes spatial features and type features into multidimensional feature vectors based on the spatial coordinates obtained by beamforming and arrival angle measurement, and the type characteristics extracted from the internal data of the RFID tags, and clusters them using a weighted clustering algorithm, thereby dividing the RFID tags into multiple logical groups;

[0009] A feedback unit connected to the signal processing unit and the beamforming unit, and dynamically adjusting the direction and width of the beam, as well as the order and area of ​​scanning based on the real-time signal parameters;

[0010] The data management unit is connected to the signal processing unit and is used to store and manage the tag data and output the reading result of the RFID tag.

[0011] As an optional implementation, the anti-collision algorithm includes:

[0012] Based on the physical location and type characteristics of the RFID tags, the RFID tags are divided into a plurality of logical groups;

[0013] Initialize the frame length N for each group of tags and broadcast the frame parameters to the tags;

[0014] After receiving the frame parameters, the RFID tag randomly selects a time slot to respond. When its response time slot matches the current frame parameters, it sends data, otherwise it enters the standby state and waits for the next frame;

[0015] In response to the end of each frame, real-time feedback information of the tag response is collected by detecting the number of collision time slots, the number of successful time slots, and the number of idle time slots;

[0016] Based on the real-time feedback information, the number of tags that are not currently recognized is calculated, and the frame length of the next frame is dynamically adjusted according to the calculation result;

[0017] In response to the collision rate being greater than or equal to a first preset threshold, increasing the frame length N;

[0018] In response to the collision rate being less than or equal to the second preset threshold, the frame length N is reduced.

[0019] As an optional implementation, the anti-collision algorithm further includes:

[0020] During the frame adjustment process, the region where the tag is located is divided into sub-regions, and the frame length is adjusted independently based on the tag response conflict situation in each sub-region;

[0021] Based on the number of unidentified tags and the response conflict rate in each sub-area, the corresponding number of time slots is allocated;

[0022] For a tag that has been successfully identified, its unique identifier is recorded, and time slot allocation for the tag that has been successfully identified is skipped in subsequent frames;

[0023] When it is detected that the unique identifications of all tags have been recorded and there is no collision time slot or no response time slot in the current frame, the identification process is terminated.

[0024] As an optional implementation, the weighted clustering algorithm includes:

[0025] The weighted K-means algorithm is used to perform clustering using the total distance metric calculated using the activation function;

[0026] Among them, the total distance metric The expression is: ;

[0027] in, For label and cluster centers The spatial distance is the difference measure of type features, and It is a weight coefficient preset according to the optical cable management requirement information.

[0028] As an optional implementation, dynamically adjusting the direction and width of the beam, and the order and area of ​​scanning based on the real-time signal parameters output by the signal processing unit includes:

[0029] The signal processing unit is used to collect the response signal characteristic data of the RFID tag in real time, including the arrival angle and the received signal strength indication;

[0030] The spatial distribution model of RFID tags is constructed, and the density peak clustering algorithm is used to identify tag-dense areas and tag-sparse areas based on the AOA and RSSI data of the tags;

[0031] Based on the spatial distribution model, the direction and width of the beam are dynamically adjusted.

[0032] As an optional implementation manner, dynamically adjusting the direction and width of the beam based on the spatial distribution model includes:

[0033] For the label-dense areas identified by the density peak clustering algorithm, the gradient descent algorithm is used to iteratively adjust the beam direction and width to maximize the label recognition rate and calculate the optimal value of the beam parameters;

[0034] Based on the beam parameters calculated by the gradient descent algorithm, the beamforming unit is controlled to adjust the amplitude and phase distribution of the antenna array, adjust the direction and width of the beam, and dynamically update according to the real-time signal parameters.

[0035] As an optional implementation manner, the dynamic updating according to the real-time signal parameter includes:

[0036] Based on the real-time signal parameters, the offset correction value of the beam direction is calculated, and the beam direction is adjusted to focus on the target area in combination with the signal strength and the arrival angle change;

[0037] After each scan, the beam parameters are continuously optimized based on the newly acquired real-time signal parameters until all RFID tags are identified.

[0038] Compared with the prior art, the beneficial effects of the present invention are as follows: through the combination of multi-antenna array and digital beamforming technology, efficient and accurate identification in high-density RFID tag scenarios is achieved. The system can dynamically adjust the beam direction and width, concentrate the radio frequency energy on the target area, and significantly improve the efficiency of tag activation and recognition; combined with real-time signal feedback and dynamic optimization strategies, the beam parameters are optimized through density peak clustering algorithm and gradient descent algorithm, and the signal collision rate and missed reading rate are reduced, thereby ensuring a high recognition rate. In addition, the present invention adopts a logical grouping strategy and an intelligent anti-collision algorithm to dynamically adjust the frame and partition management of the tag, so as to achieve efficient reading and intelligent management of multi-region and multi-type tags. The data management unit supports the classified storage and analysis of tag data, providing powerful intelligent management capabilities for complex scenarios such as logistics, warehousing, and production lines. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 A schematic diagram of a high-density RFID tag group scanning system based on a multi-antenna array and beamforming provided in an embodiment of the present invention;

[0040] Figure 2 A flowchart of an anti-collision algorithm provided for the implementation of the present invention;

[0041] Figure 3 A flow chart of dividing RFID tags into multiple logical groups is provided in an embodiment of the present invention.

[0042] Reference numerals:

[0043] 10. Multi-antenna array; 20. RFID reader / writer; 30. Beamforming unit; 40. Signal processing unit; 50. Feedback unit; 60. Data management unit. DETAILED DESCRIPTION

[0044] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0045] See also Figure 1 , Figure 1 A schematic diagram of a high-density RFID tag group scanning system based on a multi-antenna array and beamforming provided in an embodiment of the present invention, the system includes: a multi-antenna array 10, an RFID reader 20, a beamforming unit 30, a signal processing unit 40, a feedback unit 50 and a data management unit 60, wherein:

[0046] A multi-antenna array 10, for receiving a response signal from an RFID tag;

[0047] An RFID reader / writer 20, connected to the multi-antenna array 10, for transmitting an inquiry signal to the RFID tag and receiving a response signal transmitted by the multi-antenna array 10;

[0048] A beamforming unit 30 is connected to the RFID reader / writer 20 and the multi-antenna array 10, controls the amplitude and phase distribution of the multi-antenna array 10, generates a directional beam, and scans a target area containing an RFID tag;

[0049] The signal processing unit 40 is connected to the RFID reader / writer 20, receives the response signal, analyzes and distinguishes the response signals of multiple RFID tags using an anti-collision algorithm, and generates corresponding tag data and real-time signal parameters;

[0050] A feedback unit 50 connected to the signal processing unit 40 and the beamforming unit 30;

[0051] Based on the real-time signal parameters, the amplitude and phase distribution of the multi-antenna array 10 are adjusted by controlling the beamforming unit 30, and the direction and width of the beam, as well as the order and area of ​​scanning are dynamically adjusted;

[0052] The data management unit 60 is connected to the signal processing unit 40 and is used to store and manage the tag data and output the reading result of the RFID tag.

[0053] In a specific implementation, the multi-antenna array 10 is composed of a number of antenna units, which are usually arranged into a uniform linear array (ULA) or a planar array. Taking ULA as an example, the spacing between antenna units is set to half a wavelength (λ / 2), and the spacing is about 16.4 cm based on the operating frequency of 915 MHz. This arrangement helps to form good beam directivity and spatial resolution. Each antenna unit has the ability to receive the response signal of the RFID tag, and through reasonable layout and connection methods, a complete antenna array is formed.

[0054] The RFID reader 20 is connected to the multi-antenna array 10 and is responsible for transmitting an inquiry signal to the RFID tag and receiving a response signal from the tag. The reader complies with the EPCglobal Gen2 protocol standard and can activate the RFID tag in the target area and send identification and data reading instructions. Through the RF front-end circuit, the reader and the antenna array realize bidirectional signal transmission, ensuring the effective transmission and reception of the inquiry signal and the response signal.

[0055] The beamforming unit 30 is located between the RFID reader 20 and the multi-antenna array 10, and plays a role in controlling the amplitude and phase distribution of the antenna array. The unit includes hardware components such as a digital signal processor (DSP), a field programmable gate array (FPGA), and a radio frequency control module. By applying appropriate amplitude weighting and phase offset to each antenna unit, the beamforming unit 30 can generate a directional beam with a specific direction and width to achieve accurate scanning of the target area. For example, in order to point the beam to a specific angle θ, the beamforming unit 30 calculates and applies the corresponding phase offset so that the transmitted signals of each antenna unit form a coherent superposition in this direction to enhance the signal strength.

[0056] The signal processing unit 40 is connected to the RFID reader 20 and is mainly responsible for receiving and processing the response signal from the RFID tag. The unit includes components such as a high-speed analog-to-digital converter (ADC), a digital signal processor (DSP) and a memory, and has a strong data processing capability. The signal processing unit 40 first down-converts, demodulates and filters the received signal to extract the effective baseband signal. Subsequently, the anti-collision algorithm is used to parse and distinguish the response signals of multiple RFID tags to avoid data loss caused by signal conflicts. Commonly used anti-collision algorithms include time slot ALOHA, Q algorithm, etc. These algorithms effectively reduce the collision probability of tag responses through time slot randomization and frame parameter adjustment. In addition, the signal processing unit 40 also extracts the real-time signal parameters of the tag response signal, such as the received signal strength indication (RSSI), phase information and angle of arrival (AOA), which provides a basis for subsequent beam adjustment and scanning strategies.

[0057] The feedback unit 50 is connected to the signal processing unit 40 and the beamforming unit 30, and dynamically adjusts the direction and width of the beam, as well as the order and area of ​​scanning based on real-time signal parameters. When the signal processing unit 40 detects that the tag response signal in certain directions is weak, or the recognition rate is lower than a preset threshold, the feedback unit 50 calculates the required beam adjustment parameters and instructs the beamforming unit 30 to adjust the amplitude and phase distribution of the antenna array. For example, by changing the pointing angle of the beam, it can be more focused on the target area that has not been fully scanned; or by adjusting the beam width, the beam can be reduced in the tag-dense area to enhance the signal strength, and the beam can be expanded in the tag-sparse area to improve the scanning coverage.

[0058] The data management unit 60 is connected to the signal processing unit 40 and is responsible for storing and managing the acquired tag data. The unit includes a database system and data processing software, which can classify and store data such as the unique identification, type characteristics and location information of the tag. The data management unit 60 also provides a data query and output interface, which can transmit the reading results of the RFID tag to the host computer or other management system to achieve real-time monitoring and subsequent processing of the tag information. In addition, for the successfully identified tags, the data management unit 60 will update its status to avoid repeated reading in subsequent scans, thereby improving the overall efficiency of the system.

[0059] Exemplarily, first, the RFID reader 20 transmits an inquiry signal to the target area through the multi-antenna array 10 to activate the RFID tags in the area. After receiving the inquiry signal, the tag randomly selects a time slot to respond according to the requirements of the protocol and the anti-collision algorithm. The multi-antenna array 10 receives the response signal from the tag and transmits the signal to the RFID reader 20 and the signal processing unit 40. The signal processing unit 40 demodulates and filters the signal, uses the anti-collision algorithm to parse the responses of multiple tags, and extracts the tag data and real-time signal parameters.

[0060] Based on the real-time signal parameters, the feedback unit 50 determines whether the direction and width of the beam need to be adjusted. If the tag recognition rate in some areas is low, or there are more signal collisions, the feedback unit 50 will calculate new beam parameters and adjust the amplitude and phase distribution of the antenna array through the beamforming unit 30 to optimize the directivity and coverage of the beam. This dynamic adjustment process continues until all RFID tags in the target area are successfully identified or the preset scanning conditions are met.

[0061] In this way, the system of the present invention can achieve efficient and accurate tag scanning and identification in a high-density RFID tag environment. The combination of multi-antenna array and beamforming technology enables the system to flexibly control the spatial distribution of radio frequency energy and enhance the signal strength in the target area. The coordinated work of the signal processing unit and the feedback unit ensures the system's effective analysis and real-time adjustment of the tag response signal, thereby improving the success rate of tag identification and the overall performance of the system.

[0062] For example, this system is suitable for application scenarios that require efficient reading of a large number of RFID tags, such as logistics warehousing, retail inventory management, and production line tracking. Through precise beam control and intelligent anti-collision algorithms, the system can achieve fast and reliable tag recognition under complex environmental conditions, providing strong technical support for the widespread application of RFID technology.

[0063] See also Figure 2 , Figure 2 A flowchart of an anti-collision algorithm provided for the implementation of the present invention;

[0064] As an optional implementation, the anti-collision algorithm includes steps S101 to S106, wherein:

[0065] S101: Divide the RFID tags into a plurality of logical groups based on the physical locations and type characteristics of the RFID tags;

[0066] S102: Initialize the frame length N for each group of tags and broadcast the frame parameters to the tags;

[0067] S103: After receiving the frame parameters, the RFID tag randomly selects a time slot to respond, and sends data when its response time slot matches the current frame parameters, otherwise it enters a standby state and waits for the next frame;

[0068] S104: In response to the end of each frame, real-time feedback information of the tag response is collected by detecting the number of collision time slots, the number of successful time slots, and the number of idle time slots;

[0069] S105: estimating the number of currently unrecognized tags based on the real-time feedback information, and dynamically adjusting the frame length of the next frame according to the estimation result;

[0070] S106a: In response to the conflict rate being greater than or equal to the first preset threshold, increasing the frame length N;

[0071] S106b: In response to the collision rate being less than or equal to the second preset threshold, reducing the frame length N.

[0072] In the system of the present invention, in order to improve the recognition efficiency in a high-density RFID tag environment, an improved anti-collision algorithm is adopted. The algorithm is based on the analysis of the physical location and type characteristics of the RFID tags, and divides the tags into multiple logical groups, thereby effectively reducing the signal collision between tags during the communication process.

[0073] In a specific implementation, the signal processing unit 40 is first used to obtain the physical location information and type characteristics of each RFID tag. The physical location information can be obtained by measuring the angle of arrival (AOA), and the type characteristics are extracted from the data field inside the RFID tag, such as product category, manufacturing batch, etc. This information is used to divide the tags into several logical groups, each of which contains tags with similar physical locations or type characteristics.

[0074] For each logical group, the system initializes a frame length N, which is usually set according to the estimated number of tags in the group. For example, if a logical group is estimated to contain 100 tags, the initial frame length N can be set to 128. The RFID reader 20 sends the corresponding frame parameters, including the frame length N and other necessary communication parameters, to all tags in the logical group via broadcast.

[0075] After receiving the frame parameters, the RFID tag randomly selects a time slot in the current frame to respond according to the anti-collision protocol (for example, the slotted ALOHA protocol). If the time slot selected by the tag matches the current frame parameters of the reader, the tag will send its own identification information; if not, the tag will enter the standby state and wait for the start of the next frame.

[0076] After each frame ends, the signal processing unit 40 collects real-time feedback information of the tag response by counting the number of successfully received time slots (S), the number of collision time slots (C), and the number of idle time slots (I). Among them, a successful time slot means that only one tag successfully sends data in the time slot; a collision time slot means that multiple tags send data at the same time, resulting in signal conflict; and an idle time slot is a time slot where no tag sends data. The total frame length satisfies the relationship:

[0077]

[0078] Based on these real-time feedback information, the system uses an anti-collision algorithm to estimate the number of tags that are not currently recognized. Commonly used estimation algorithms such as the Schoute estimation algorithm can be estimated using the following formula:

[0079]

[0080] in, is the estimated number of unidentified tags, is the number of collision time slots.

[0081] Based on the estimation results, the system dynamically adjusts the frame length N of the next frame to optimize communication efficiency. ) is greater than or equal to the first preset threshold (for example, 0.4), indicating that the collision of the tag response is more serious. The system will increase the frame length N, for example, double the current frame length, so that the tag can be randomly selected within a larger time slot range to reduce the collision probability.

[0082] If the collision rate is less than or equal to the second preset threshold (for example, 0.1), it means that the tag responses are relatively scattered and the time slot resources may not be fully utilized. The system reduces the frame length N, for example, halves the current frame length, to improve the time slot utilization and speed up the recognition speed.

[0083] By dynamically adjusting the frame length, the system can adaptively optimize communication parameters and improve the recognition efficiency of RFID tags. This method is particularly suitable for scenarios where the number of tags changes dynamically or the tags are unevenly distributed, and can maintain high recognition performance under different tag densities.

[0084] In addition, in order to further improve the stability and response speed of the system, the maximum and minimum values ​​of the frame length can be set during the frame length adjustment process to prevent the frame length from being too large, which will increase the delay, or too small, which will prevent the collision from being effectively reduced.

[0085] In this way, the anti-collision algorithm of the present invention performs logical grouping based on the physical location and type characteristics of the tags, and dynamically adjusts the frame length in combination with real-time feedback information, thereby effectively reducing the collision probability of tag responses and significantly improving the recognition efficiency in a high-density RFID tag environment.

[0086] As an optional implementation, the anti-collision algorithm further includes:

[0087] During the frame adjustment process, the region where the tag is located is divided into several sub-regions, and the frame length is adjusted independently based on the tag response conflict situation in each sub-region;

[0088] Based on the number of unidentified tags and the response conflict rate in each sub-area, the corresponding number of time slots is allocated;

[0089] For a tag that has been successfully identified, its unique identifier is recorded, and time slot allocation for the tag that has been successfully identified is skipped in subsequent frames;

[0090] When it is detected that the unique identifications of all tags have been recorded and there is no collision time slot or no response time slot in the current frame, the identification process is terminated.

[0091] In the system of the present invention, in order to further improve the recognition efficiency in a high-density RFID tag environment, the anti-collision algorithm not only divides the tags into multiple logical groups based on the physical location and type characteristics of the tags, but also further divides the area where the tags are located into several sub-areas during the frame adjustment process. In this way, the system can independently adjust the frame length and communication parameters based on the tag response collision situation in each sub-area, thereby more finely controlling the tag recognition process.

[0092] In the specific implementation, the spatial location information of each RFID tag is first obtained by using the signal processing unit 40, for example, through parameters such as angle of arrival (AOA) and received signal strength indication (RSSI), to build a spatial distribution model of the tag. According to the model, the entire tag area is divided into several sub-areas. The division method can be grid division based on spatial coordinates, or dynamic division based on tag density. For example, a tag-dense area is divided into smaller sub-areas for more precise control.

[0093] In each sub-area, the system counts the number of unrecognized tags and the response collision rate. The number of unrecognized tags can be calculated by the difference between the number of accumulated recognized tags and the total number of tags initially estimated, or estimated by the number of real-time collision time slots. The response collision rate is obtained by the ratio of the number of collision time slots to the total number of time slots in the current frame.

[0094] Based on the above statistical information, the system independently adjusts the frame length N for each sub-area, that is, allocates the corresponding number of time slots. For sub-areas with a large number of unidentified tags and a high response conflict rate, the system increases its frame length to reduce the collision probability of tag responses. Conversely, for sub-areas where most tags have been successfully identified and the response conflict rate is low, the system can appropriately reduce the frame length to improve communication efficiency.

[0095] For successfully identified tags, the system records their unique identifiers (such as EPC codes) and skips allocating time slots to these identified tags in subsequent communication frames. Specifically, after receiving a command from the reader, the successfully identified tags update their status to "identified" or "silent" and no longer participate in the response in subsequent frames. This mechanism effectively reduces unnecessary communication overhead and prevents identified tags from interfering with the communication of unidentified tags.

[0096] When the system detects that the unique identification of all tags has been recorded and there is no collision time slot or no response time slot in the current frame, it means that all tags have been successfully identified and no new tags have entered the communication range, and the system can terminate the identification process. This termination condition setting ensures that the system can stop scanning in time after completing the identification of all tags, saving resources.

[0097] Exemplarily, the target area is divided into several sub-areas according to the spatial distribution of the tags, and the number and density of tags in each sub-area may be different.

[0098] For each sub-region, the frame length N is initialized and the frame parameters are broadcast to the tags in the sub-region. The initial value of the frame length can be set according to the estimated number of tags in the sub-region.

[0099] After receiving the frame parameters, the RFID tag randomly selects a time slot in the current frame to respond according to the anti-collision algorithm. When the selected time slot matches the current frame parameters, it sends its own identification information; otherwise, it enters the standby state and waits for the next frame.

[0100] The signal processing unit 40 receives and analyzes the response signal of the tag, counts the number of successful time slots, the number of collision time slots and the number of idle time slots, and obtains real-time feedback information.

[0101] Based on the real-time feedback information, the number of unrecognized tags and the response conflict rate in each sub-area are estimated. According to the estimation results, the frame length N of each sub-area is adjusted independently. For sub-areas with higher conflict rates, the frame length is increased; for sub-areas with lower conflict rates, the frame length is reduced.

[0102] For successfully identified tags, their unique identifiers are recorded and no time slots are allocated to them in subsequent frames. Identified tags enter a "silent" state and no longer participate in communications to avoid interfering with the responses of unidentified tags.

[0103] When the tags in all sub-areas have been successfully identified, that is, the unique identifiers of all tags have been recorded, and there are no collision time slots or unresponse time slots in the current frame, the system determines that the identification process is completed and terminates the communication.

[0104] In this way, the system can effectively reduce the collision probability of tag responses and improve recognition efficiency in a high-density tag environment. At the same time, the frame length of each sub-area is adjusted independently, so that system resources are optimized and the waste of resources or increased conflicts caused by uniformly adjusting the frame length is avoided.

[0105] In addition, during the entire recognition process, the system can dynamically adjust the division of sub-regions and the setting of frame parameters based on real-time data feedback. For example, after the tag recognition in a sub-region is completed, the system can re-divide the adjacent sub-regions or merge resources to further improve the recognition efficiency.

[0106] In this way, the present invention significantly improves the recognition efficiency of RFID tags and the resource utilization of the system by introducing the mechanism of sub-area division and independent frame length adjustment in the anti-collision algorithm and combining it with the management strategy of successfully identified tags. This method is particularly suitable for application scenarios with high tag density and uneven distribution.

[0107] See also Figure 3 , Figure 3 A flowchart of dividing an RFID tag into multiple logical groups is provided in an embodiment of the present invention; as an optional implementation, the dividing the RFID tag into multiple logical groups based on the physical location and type characteristics of the RFID tag includes steps S201 to S204, wherein:

[0108] S201: using the signal processing unit to collect signal feature data of each RFID tag, wherein the signal feature data includes spatial coordinates obtained by beamforming and angle of arrival measurement, and type characteristics extracted from internal data of the RFID tag;

[0109] S202: Encoding the spatial coordinates and type characteristics into numerical features to form a multi-dimensional feature vector;

[0110] S203: Based on the multidimensional feature vector, a weighted clustering algorithm is used to cluster the RFID tags, wherein the spatial features and the type features are weightedly fused according to preset weights, and the weights are determined according to the requirements of the optical cable management scenario;

[0111] S204: Divide the RFID tags into a plurality of logical groups according to the clustering result.

[0112] As an optional implementation, the weighted clustering algorithm includes:

[0113] The weighted K-means algorithm is used to perform clustering using the total distance metric calculated using the activation function;

[0114] Among them, the total distance metric The expression is: ;

[0115] in, For label and cluster centers The spatial distance is the difference measure of type features, and It is a weight coefficient preset according to the optical cable management requirement information.

[0116] In the system of the present invention, in order to effectively manage and identify high-density RFID tags, the RFID tags are divided into multiple logical groups based on their physical locations and type characteristics.

[0117] In a specific implementation, the signal processing unit 40 is used to collect signal feature data of each RFID tag. The signal feature data includes spatial coordinates obtained by beamforming and angle of arrival (AOA) measurement, as well as type characteristics extracted from the internal data of the RFID tag. The spatial coordinates can be obtained by using a phase difference measurement method of a multi-antenna array. Specifically, the signal processing unit 40 performs coherent processing on the received tag response signal, and uses the MUSIC (Multiple Signal Classification) algorithm or the ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques) algorithm to accurately estimate the angle of arrival of the tag. Combined with the geometric structure of the antenna array, the spatial position coordinates (x, y, z) of each tag are calculated.

[0118] The type characteristics are extracted from the data stored inside the RFID tag. For example, the electronic product code (EPC) of the tag may contain information such as the category, model, and batch of the item. The signal processing unit 40 reads these type characteristics during the decoding of the tag response signal, providing a basis for subsequent data processing.

[0119] The spatial coordinates and type characteristics are encoded as numerical features to form a multidimensional feature vector. The spatial coordinates are directly used as numerical features in the calculation. For type characteristics, if it is category data, such as item category, model, etc., One-Hot encoding or Label Encoding can be used to convert it into numerical form. For example, assuming there are three item categories A, B, and C, they can be encoded as (1,0,0), (0,1,0), and (0,0,1). If the type characteristic is numerical data, such as weight, size, etc., it can be used directly.

[0120] The formed multi-dimensional feature vector is expressed as:

[0121]

[0122] in, is the feature vector of the i-th label, represents the x, y, and z coordinate components of the i-th RFID tag in the spatial coordinate system, The value encoded for the type attribute, The dimension representing the type feature, that is, the number of numerical features contained in the type feature of the RFID tag after being encoded. For each tag, its type feature may include multiple attributes, such as product category, model, batch, etc. These attributes are encoded to form a A vector of numeric features.

[0123] For example, if the label contains three type features: category, model, and batch, after encoding, the type feature vector may be ,at this time .

[0124] Based on the multidimensional feature vector, a weighted clustering algorithm is used to cluster the RFID tags. In this embodiment, a weighted K-means clustering algorithm is selected. The algorithm assigns different weights to each feature dimension to reflect its importance in clustering. The weights are determined based on the needs of the optical cable management scenario. For example, in some scenarios, the spatial location may be more important than the type characteristics, and vice versa.

[0125] The specific steps of the weighted K-means algorithm are as follows:

[0126] K initial cluster centers are randomly selected or set based on prior knowledge. Each cluster center is also a vector with the same dimension as the multidimensional feature vector.

[0127] For each multidimensional feature vector, calculate its weighted distance to each cluster center. The total distance metric The expression is:

[0128]

[0129] in, is the spatial distance between label i and cluster center j, calculated using Euclidean distance:

[0130]

[0131] in, are the x, y, and z coordinate components of the jth cluster center in the spatial coordinate system;

[0132] As the difference measure of type features, Manhattan distance or Euclidean distance can be used:

[0133]

[0134] in, is the feature value of the i-th tag in the k-th type feature dimension. The type feature is encoded from the internal data of the RFID tag (such as the item category, model, batch, etc. in the EPC), and is the component of the feature vector formed by One-Hot encoding or digitization; Represents the feature value of the j-th cluster center on the k-th type feature dimension; and is the preset weight coefficient, satisfying , set according to the cable management requirements. For example, if the spatial feature is more important, you can set .

[0135] Each label is assigned to the cluster to which the nearest cluster center belongs based on the calculated total distance metric.

[0136] For each cluster, calculate its new cluster center, which is the weighted average of all feature vectors in the cluster.

[0137] Repeat the above steps until the cluster center no longer changes or the maximum number of iterations is reached.

[0138] According to the clustering results, the RFID tags are divided into multiple logical groups. Each logical group contains tags with close spatial locations and similar type characteristics. In the subsequent anti-collision algorithm and beamforming process, the communication parameters and scanning strategies can be adjusted independently for each logical group.

[0139] In specific applications, such as in the optical cable management scenario, different types of optical cables need to be distinguished and managed according to their physical locations. Through the above weighted clustering algorithm, RFID tags can be effectively grouped according to the type of optical cable and the installation location. In this way, when reading the tag information, the beam direction and width can be adjusted for specific logical groups to improve the reading efficiency.

[0140] In addition, the determination of the weight coefficient can be adjusted according to actual needs. If you need to focus on a certain feature, you can increase the weight of the feature appropriately. For example, when you need to accurately locate the position of the label, you can increase the weight of the spatial feature; when you need to distinguish different types of labels, you can increase the weight of the type feature.

[0141] In order to improve the efficiency and accuracy of clustering, the feature vectors can be standardized before clustering to make the data of each dimension on the same order of magnitude, so as to avoid a certain feature dominating the distance calculation due to a large numerical range.

[0142] Through the above method, the present invention realizes effective logical grouping of tags based on the physical location and type characteristics of RFID tags. This logical grouping method based on weighted clustering algorithm makes full use of the multi-dimensional characteristics of tags and improves the recognition efficiency and accuracy of the system in a high-density tag environment.

[0143] In subsequent system operations, the parameters of the anti-collision algorithm can be optimized for the divided logical groups, such as independently adjusting the frame length N, time slot allocation, etc. At the same time, during beamforming, the direction and width of the beam can be accurately adjusted according to the spatial position of the logical group to focus on scanning a specific area.

[0144] In this way, the present invention utilizes the signal feature data collected by the signal processing unit and adopts a weighted clustering algorithm to divide the RFID tags into multiple logical groups, which fully utilizes the spatial and type characteristics of the tags, improves the operability and practicality of the system, and provides strong technical support for efficient RFID tag group scanning.

[0145] As an optional implementation, dynamically adjusting the direction and width of the beam, and the order and area of ​​scanning based on the real-time signal parameters output by the signal processing unit includes:

[0146] The signal processing unit is used to collect the response signal characteristic data of the RFID tag in real time, including the arrival angle and the received signal strength indication;

[0147] The spatial distribution model of RFID tags is constructed, and the density peak clustering algorithm is used to identify tag-dense areas and tag-sparse areas based on the AOA and RSSI data of the tags;

[0148] Based on the spatial distribution model, the direction and width of the beam are dynamically adjusted.

[0149] In the system of the present invention, in order to improve the recognition efficiency and accuracy of RFID tags, the direction and width of the beam, as well as the order and area of ​​scanning are dynamically adjusted based on the real-time signal parameters output by the signal processing unit 40.

[0150] In the specific implementation, first, the signal processing unit 40 is used to collect the response signal characteristic data of the RFID tag in real time. The characteristic data includes the angle of arrival (AOA) and the received signal strength indicator (RSSI) of each tag. AOA is obtained by receiving the response signal of the tag through a multi-antenna array, and using a high-resolution angle of arrival estimation algorithm, such as the MUSIC (Multiple Signal Classification) algorithm or the ESPRIT (Estimation of Signal Parameters via Rotational Invariance Techniques) algorithm, to achieve accurate measurement of the incident direction of the tag signal. RSSI is directly obtained by measuring the strength of the received signal, which reflects the distance between the tag and the receiving antenna and the attenuation of the signal propagation path.

[0151] Next, the spatial distribution model of RFID tags is constructed based on the AOA and RSSI data collected in real time. To this end, the density peak clustering algorithm (Clustering by Fast Search and Find of Density Peaks, referred to as the DPC algorithm) is used to perform cluster analysis on the tags. This algorithm does not require the number of clusters to be specified in advance, and can automatically identify the clustered areas and sparse areas of the data based on the density distribution of the data, thereby identifying the areas with dense tags and the areas with sparse tags.

[0152] For example, for each label, the local density is calculated according to its proximity in the feature space. The calculation formula is:

[0153]

[0154] in, is the distance between label i and label p, is the cutoff distance, is a step function, when hour, ,otherwise .distance The Euclidean distance can be calculated using the feature vector composed of AOA and RSSI:

[0155]

[0156] in, and are the AOA of label i and label p respectively, and are the received signal strengths of tag i and tag p respectively.

[0157] For each label, calculate the minimum distance between it and the labels with higher density :

[0158]

[0159] For the label with the highest density, Take all Maximum value.

[0160] according to and Draw a decision diagram and choose and high The labels are used as cluster centers, i.e., dense areas of identification labels.

[0161] Assign the remaining labels to the cluster to which the cluster center closest to them belongs to complete the clustering of the labels.

[0162] Furthermore, based on the constructed RFID tag spatial distribution model, the system dynamically adjusts the direction and width of the beam.

[0163] Exemplarily, for the identified tag-dense area, the average AOA of the tags in the area is calculated as the target beam direction. The beamforming unit 30 adjusts the phase and amplitude distribution of the multi-antenna array according to the direction, so that the main lobe of the beam points to the tag-dense area, thereby enhancing the signal reception strength in the area and improving the tag recognition rate.

[0164] Adjust the beam width according to the spatial range of the tag-dense area. In the tag-dense area, appropriately reduce the beam width to concentrate the energy and improve the signal-to-noise ratio; in the tag-sparse area, expand the beam width to cover a larger space and improve the scanning efficiency.

[0165] The scanning order is dynamically adjusted according to the density of tags in each area and the recognition priority. Areas with dense tags are scanned first to ensure that a large number of tags in these areas can be quickly recognized.

[0166] Exemplarily, the signal processing unit 40 continuously receives the response signal of the RFID tag, extracts the AOA and RSSI data in real time, and pre-processes the data, such as normalization and filtering, to reduce the influence of noise. The collected AOA and RSSI data are regularly (for example, every several seconds) subjected to density peak cluster analysis to update the spatial distribution model of the RFID tag. According to the latest spatial distribution model, the beam direction and width parameters that need to be adjusted are calculated. For each identified tag-dense area, its center position and range are calculated as the basis for beam adjustment. The beam forming unit 30 receives the new beam parameters, adjusts the amplitude and phase distribution of the antenna array, and realizes real-time adjustment of the beam. Specifically, by changing the excitation phase and amplitude weight of each antenna unit, the beam is pointed in the target direction to meet the specific beam width requirements. The system scans each area in an optimized scanning order. For tag-dense areas, the number of scans or the dwell time can be increased to ensure that the tag is successfully identified. After each scan, the signal processing unit 40 analyzes the identification status and signal quality of the tag. If it is found that the recognition rate in certain areas is still low, the system can further adjust the beam parameters or take other compensatory measures, such as increasing the transmission power.

[0167] In this way, the system of the present invention can dynamically adjust the direction and width of the beam, as well as the order and area of ​​scanning according to the real-time spatial distribution of the RFID tags. This adaptive beam control strategy makes full use of the spatial characteristics and signal characteristics of the RFID tags, and improves the recognition efficiency and reliability in a high-density tag environment.

[0168] In addition, in order to improve the response speed and computing efficiency of the system, acceleration technology can be used in the density peak clustering algorithm, such as KD tree or Ball tree to accelerate distance calculation. For large-scale labeled data sets, parallel computing or distributed computing methods can be used to improve the efficiency of clustering analysis.

[0169] In addition, in the implementation of beamforming, Digital Beamforming (DBF) technology can be used to achieve precise control of the antenna array through a high-speed digital signal processor (DSP) or field programmable gate array (FPGA). DBF technology has the characteristics of high flexibility and high precision, and is suitable for the needs of real-time adjustment of beam parameters.

[0170] In the signal processing unit 40, a machine learning algorithm can be combined to perform pattern recognition and classification on the tag response signal to further improve the accuracy of tag recognition. For example, a deep neural network can be used to learn the tag signal characteristics, identify complex signal patterns, and adapt to complex environments such as multipath propagation and noise interference.

[0171] In this way, the present invention constructs a spatial distribution model of RFID tags by using a density peak clustering algorithm based on the real-time signal parameters output by the signal processing unit 40, and dynamically adjusts the direction and width of the beam, as well as the order and area of ​​scanning, thereby realizing efficient identification and management of high-density RFID tags. The system has the advantages of high precision, high efficiency and high reliability, and is suitable for application scenarios such as warehousing logistics and production line tracking that require rapid identification of a large number of RFID tags.

[0172] As an optional implementation manner, dynamically adjusting the direction and width of the beam based on the spatial distribution model includes:

[0173] For the label-dense areas identified by the density peak clustering algorithm, the gradient descent algorithm is used to iteratively adjust the beam direction and width to maximize the label recognition rate and calculate the optimal value of the beam parameters;

[0174] Based on the beam parameters calculated by the gradient descent algorithm, the beamforming unit is controlled to adjust the amplitude and phase distribution of the antenna array, adjust the direction and width of the beam, and dynamically update according to the real-time signal parameters.

[0175] As an optional implementation manner, the dynamic updating according to the real-time signal parameter includes:

[0176] Based on the real-time signal parameters, the offset correction value of the beam direction is calculated, and the beam direction is adjusted to focus on the target area in combination with the signal strength and the arrival angle change;

[0177] After each scan, the beam parameters are continuously optimized based on the newly acquired real-time signal parameters until all RFID tags are identified.

[0178] In the system of the present invention, in order to further improve the recognition efficiency of RFID tags, based on the spatial distribution model, a gradient descent algorithm is used to optimize and adjust the beam direction and width for the tag-dense area identified by the density peak clustering algorithm.

[0179] In the specific implementation, first, the system uses the density peak clustering algorithm to analyze the real-time signal parameters of RFID tags and build a spatial distribution model. By clustering the data such as the angle of arrival (AOA) and received signal strength indication (RSSI) of the tags, the tag-dense areas are identified. These areas often contain a large number of tags to be identified, so they need to be paid special attention.

[0180] For each tag-dense area, the system aims to maximize the tag recognition rate and establishes an objective function. The beam direction and width are set as adjustable parameters, and the objective function can be expressed as the relationship between the tag recognition rate and the beam parameters:

[0181]

[0182] Among them, A is the recognition rate, It represents the functional relationship between the recognition rate and the beam pointing angle θ and the beam width Δθ. θ is the central pointing direction angle of the beam, and Δθ is the main lobe width of the beam (i.e. the angle range covered by the beam). When θ is close to the average incident direction of the target tag dense area, and Δθ can reasonably cover the area, The corresponding recognition rate will be improved. In order to find the beam parameter combination that maximizes the tag recognition rate, the system uses the gradient descent algorithm for iterative optimization.

[0183] Exemplarily, the specific steps of the gradient descent algorithm are as follows:

[0184] According to the spatial distribution model, the initial beam direction and width are selected so that the beam initially covers the tag-dense area.

[0185] Under the current beam parameters, perform a tag scan, record the number of tags successfully identified, and calculate the recognition rate.

[0186] By making slight perturbations to the beam direction and width, the changes in the recognition rate are measured respectively, and the partial derivative of the objective function with respect to the beam parameters, i.e., the gradient, is calculated:

[0187]

[0188]

[0189] in, These are small changes introduced when finding partial derivatives, which are used to define the instantaneous rate of change of the function to the corresponding parameters. Based on the gradient information and the set learning rate, the beam direction and width are updated:

[0190]

[0191]

[0192] Respectively represent the beam pointing direction angle before and after the update, Respectively represent the beam width before and after the update, is the learning rate, which is used to control the step size of parameter update.

[0193] Repeat the above steps until the recognition rate reaches a preset maximum value or the change in beam parameters approaches zero.

[0194] In each iteration, the system controls the beamforming unit 30 to adjust the amplitude and phase distribution of the antenna array according to the new beam parameters. Specifically, the beamforming unit 30 calculates the phase offset and amplitude weighting required for each antenna element according to the updated beam direction angle. Phase offset The calculation formula is:

[0195]

[0196] in, is the antenna unit number, is the antenna spacing, is the working wavelength.

[0197] Amplitude weighting is based on the beam width Determine and use appropriate weighting functions (such as Chebyshev weighting) to control the main lobe width and side lobe level of the beam.

[0198] During the entire optimization process, the system monitors changes in signal parameters in real time, including RSSI and AOA, etc. When a change in the environment is detected (such as tag movement or channel status change), the system restarts the gradient descent algorithm and continuously optimizes the beam parameters to ensure that the recognition rate remains at the optimal level.

[0199] Exemplarily, the specific implementation of dynamic update may be:

[0200] According to the collected real-time signal parameters, calculate the angle that the beam direction needs to be adjusted. Assume that in the mth iteration, the beam direction angle is , the received tag signal is concentrated on If the offset correction value MV is near, then:

[0201]

[0202] The system updates the beam direction angle accordingly:

[0203]

[0204] In addition, the system not only considers the tag's AOA, but also combines RSSI information to evaluate signal quality. For areas with weak signal strength but high AOA concentration, the beam width may need to be adjusted to increase coverage; for areas with strong signal strength but dispersed AOA, the beam shape may need to be optimized to improve recognition rate.

[0205] After each scan, the system will re-evaluate the effectiveness of the beam parameters based on the latest signal parameters. If the recognition rate drops or the environment changes, the parameters will be adjusted immediately. This process is a real-time closed-loop control that ensures that the system always works in the best condition.

[0206] It should be noted that the learning rate determines the convergence speed and stability of the gradient descent algorithm. If the learning rate is too large, the algorithm will diverge, and if it is too small, the convergence speed will be too slow. Usually, an appropriate learning rate is determined through experiments, or an adaptive learning rate algorithm (such as the Adam algorithm) is used.

[0207] It should be noted that the gradient descent algorithm may fall into a local optimum. To this end, multiple attempts can be made on the initial parameter selection, or combined with other optimization algorithms (such as simulated annealing, genetic algorithm) for global search.

[0208] It should be noted that in order to avoid excessive adjustment of the beam direction and width, which affects the stability of the system, physical limits on the beam parameters can be set. For example, the range of change of the beam direction is limited to the scanning angle of the antenna array, and the beam width cannot be less than the resolution limit of the antenna array.

[0209] Thus, by adopting the gradient descent algorithm, the system of the present invention can dynamically optimize the direction and width of the beam according to real-time signal feedback, so that it is always focused on the target area. This method aims to maximize the tag recognition rate, iteratively adjusts the beam parameters, and ensures the best recognition effect in the tag-dense area. Combined with the precise control of the beamforming unit 30, the system can quickly respond to environmental changes and continuously improve the recognition efficiency of RFID tags.

[0210] It should be understood that determining B based on A does not mean determining B only based on A. B can also be determined based on A and / or other information.

[0211] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0212] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0213] The preferred embodiments of the present invention disclosed above are only used to help explain the present application. The preferred embodiments do not describe all the details in detail, nor do they limit the present application to specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present application, so that those skilled in the art can understand and use the present application well. The present application is limited only by the claims and their full scope and equivalents.

Claims

1. A high-density RFID tag group scanning system based on multi-antenna array and beamforming, characterized in that: include: A multi-antenna array for receiving a response signal from an RFID tag; An RFID reader / writer, connected to the multi-antenna array, configured to transmit an inquiry signal to the RFID tag and receive a response signal transmitted by the multi-antenna array; A beamforming unit connected to the RFID reader and the multi-antenna array, controlling the amplitude and phase distribution of the multi-antenna array, generating a directional beam, and scanning a target area containing an RFID tag; A signal processing unit connected to the RFID reader / writer receives and parses response signals of multiple RFID tags, parses and distinguishes the response signals of multiple RFID tags using an anti-collision algorithm, and generates tag data and real-time signal parameters, wherein the signal processing unit encodes spatial features and type features into multidimensional feature vectors based on the spatial coordinates obtained by beamforming and arrival angle measurement, and the type characteristics extracted from the internal data of the RFID tags, and clusters them using a weighted clustering algorithm, thereby dividing the RFID tags into multiple logical groups; Each logical group includes tags with close spatial positions and similar type characteristics; in the process of the anti-collision algorithm and the beamforming, the communication parameters and the scanning strategy are independently adjusted for each logical group; For each logical group, based on the number of tags in the group, initialize the frame length N for each group of tags and broadcast the frame parameters to the tags; In response to the end of each frame, real-time feedback information of the tag response is collected by detecting the number of collision time slots, the number of successful time slots, and the number of idle time slots; Based on the real-time feedback information, the number of tags that are not currently identified is calculated, and the frame length of the next frame of the logical group is dynamically adjusted according to the calculation result; A feedback unit connected to the signal processing unit and the beamforming unit, and dynamically adjusting the direction and width of the beam, as well as the order and area of ​​scanning based on the real-time signal parameters; The data management unit is connected to the signal processing unit and is used to store and manage the tag data and output the reading result of the RFID tag.

2. The high-density RFID tag group scanning system based on multi-antenna array and beamforming according to claim 1 is characterized in that: The anti-collision algorithm includes: Based on the physical location and type characteristics of the RFID tags, the RFID tags are divided into a plurality of logical groups; After receiving the frame parameters, the RFID tag randomly selects a time slot to respond. When its response time slot matches the current frame parameters, it sends data, otherwise it enters the standby state and waits for the next frame; In response to the collision rate being greater than or equal to a first preset threshold, increasing the frame length N; In response to the collision rate being less than or equal to the second preset threshold, the frame length N is reduced.

3. The high-density RFID tag group scanning system based on multi-antenna array and beamforming according to claim 2 is characterized in that: The anti-collision algorithm also includes: During the frame adjustment process, the region where the tag is located is divided into sub-regions, and the frame length is adjusted independently based on the tag response conflict situation in each sub-region; Based on the number of unidentified tags and the response conflict rate in each sub-area, the corresponding number of time slots is allocated; For a tag that has been successfully identified, its unique identifier is recorded, and time slot allocation for the tag that has been successfully identified is skipped in subsequent frames; When it is detected that the unique identifications of all tags have been recorded and there is no collision time slot or no response time slot in the current frame, the identification process is terminated.

4. The high-density RFID tag group scanning system based on multi-antenna array and beamforming according to claim 3 is characterized in that: The weighted clustering algorithm includes: The weighted K-means algorithm is used to perform clustering using the total distance metric calculated using the activation function; Among them, the total distance metric The expression is: ; in, For label and cluster centers The spatial distance is the difference measure of type features, and It is a weight coefficient preset according to the optical cable management requirement information.

5. The high-density RFID tag group scanning system based on multi-antenna array and beamforming according to claim 4 is characterized in that: Based on the real-time signal parameters output by the signal processing unit, the direction and width of the beam, as well as the order and area of ​​scanning are dynamically adjusted including: The signal processing unit is used to collect the response signal characteristic data of the RFID tag in real time, including the arrival angle and the received signal strength indication; The spatial distribution model of RFID tags is constructed, and the density peak clustering algorithm is used to identify tag-dense areas and tag-sparse areas based on the AOA and RSSI data of the tags; Based on the spatial distribution model, the direction and width of the beam are dynamically adjusted.

6. The high-density RFID tag group scanning system based on multi-antenna array and beamforming according to claim 5 is characterized in that: Based on the spatial distribution model, dynamically adjusting the direction and width of the beam includes: For the label-dense areas identified by the density peak clustering algorithm, the gradient descent algorithm is used to iteratively adjust the beam direction and width to maximize the label recognition rate and calculate the optimal value of the beam parameters; Based on the beam parameters calculated by the gradient descent algorithm, the beamforming unit is controlled to adjust the amplitude and phase distribution of the antenna array, adjust the direction and width of the beam, and dynamically update according to the real-time signal parameters.

7. The high-density RFID tag group scanning system based on multi-antenna array and beamforming according to claim 6 is characterized in that: The dynamic updating according to the real-time signal parameters comprises: Based on the real-time signal parameters, the offset correction value of the beam direction is calculated, and the beam direction is adjusted to focus on the target area in combination with the signal strength and the arrival angle change; After each scan, the beam parameters are continuously optimized based on the newly acquired real-time signal parameters until all RFID tags are identified.

Images