A secret carrier intelligent storage system fusing RFID and 3D printing

Abstract

The application discloses a secret carrier intelligent storage system fusing RFID and 3D printing, which comprises a cabinet body, an image acquisition device, an RFID accurate identification subsystem, an intelligent storage management platform and a 3D printing customized storage structure frame installed in the cabinet body; the RFID accurate identification subsystem sends the identification information collected through an antenna array to the platform, a platform digital twin mapping module constructs a three-dimensional visual model and binds information; an intelligent matching module adopts a shape recognition model to output size and shape data and inputs the data into a built-in genetic algorithm solver to output a target storage position. The application solves the problems of low space utilization of existing fixed grid openings and serious crosstalk of label signals caused by dense storage of multiple articles.

Description

Technical Field

[0001] This invention relates to the field of secure carrier management, and more specifically, to an intelligent secure carrier storage system that integrates RFID and 3D printing. Background Technology

[0002] Intelligent storage devices for classified media are primarily used for the secure storage, on-site monitoring, and traceability of classified documents, electronic media, and other physical carriers containing confidential information. As government agencies, research institutions, and large enterprises increase their requirements for the full lifecycle management of classified media, traditional management methods relying on manual registration and inventory are no longer sufficient to meet the needs of frequent access, precise location, and accountability. Therefore, combining RFID, identity authentication, cabinet control, and management platforms has become an important development direction for intelligent storage and management devices for classified media.

[0003] In the prior art, Chinese patent CN104318360A, entitled "Comprehensive Management System for Classified Carriers," discloses a management scheme including a classified carrier database, a classified carrier management and control software system, a label and QR code system, and external auxiliary hardware. This scheme uses a database and QR code identification to achieve classified carrier classification management, daily operation records, and query output, reducing the manual workload in classified carrier management. Chinese patent CN117557086A, entitled "Classified Carrier Supervision Method, Device, Equipment, and Readable Storage Medium," discloses a scheme for risk prediction of unreturned classified carriers. This scheme generates supervision prompts based on classified carrier lending information and the historical information of the lending responsible party, improving the timeliness of supervision after classified carrier lending.

[0004] However, the aforementioned existing technologies primarily focus on optimizing upper-layer data flow and software business logic, neglecting the complex electromagnetic and mechanical environment of the underlying physical execution space. Currently, security equipment in the industry faces several extremely hidden and challenging physical pain points in actual production and long-term operation. Existing equipment generally uses fixed-size molded metal or engineering plastic slots, which makes small carriers occupy too much space and irregularly shaped carriers difficult to fit. If multiple carriers are forcibly mixed in the same fixed slot, the stacked RFID tags are prone to causing severe electromagnetic crosstalk and multipath reflection, leading to large-scale missed reads in dense access scenarios. At the same time, the relatively heavy metal-encapsulated carriers, when placed statically on fixed support components in a closed environment with temperature fluctuations, are prone to extremely slow mechanical creep and deformation of the support structure, causing misalignment of the spatial relative positions of internal RF devices. In addition, when some electronic security carriers with independent power supplies or those that are already in operation are first stored in the device, their surfaces often accumulate significant residual heat. Due to the extremely high thermal resistance of the enclosed space, the local transient high temperature will directly change the physical impedance of the surrounding radio frequency circuits, causing extremely unstable radio frequency scanning blind zones in the early stages of storage, leading to frequent false alarms from the system. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide an intelligent storage system for confidential carriers that integrates RFID and 3D printing, so as to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: A smart storage system for secure carriers that integrates RFID and 3D printing includes a cabinet, an image acquisition device, an RFID precision identification subsystem, a smart storage management platform, and a 3D-printed custom storage structure frame installed inside the cabinet. The image acquisition device is connected to the intelligent storage management platform; the 3D printed custom storage structure is equipped with multiple independent storage compartments, each of which is used to place a confidential carrier; The RFID precision identification subsystem includes radio frequency tags, an antenna array, and a multi-channel reader. The radio frequency tags are set on RFID clips, which are connected to a security carrier via a retractable pull cord. Each RFID clip corresponds to a dedicated slot, which is located in a corresponding storage compartment. The antenna array is distributed in the storage compartments of the 3D printed customized storage structure frame. The multi-channel reader is connected to the antenna array and collects the identification information and received signal strength indication value of the radio frequency tags and sends them to the intelligent storage management platform. The intelligent storage management platform includes a digital twin mapping module and an intelligent matching module. The digital twin mapping module constructs a three-dimensional visualization model of the 3D-printed customized storage structure and binds the identification information to the corresponding storage compartment in the three-dimensional visualization model. The intelligent matching module uses a shape recognition model to acquire image data to be stored in the secure carrier using the image acquisition device, inputs the image data into the shape recognition model, outputs the size and shape data of the secure carrier to be stored, and inputs the size and shape data into the genetic algorithm solver built into the intelligent matching module to output the target storage compartment.

[0007] Specifically, the training process of the morphology recognition model includes: Obtain a set of sample images of the confidential carrier and manually label them with corresponding actual size and shape tags to construct a training set; The training set is input into an initial deep convolutional neural network for feature extraction and forward propagation calculation, and the predicted size and shape data are output. Calculate the loss function value between the predicted size and shape data and the actual size and shape label; The backpropagation algorithm is used to calculate the gradient of the loss function value with respect to the weights of each layer node in the deep convolutional neural network, and the weights of each layer node are updated until the loss function value is less than the set convergence threshold, thus completing the training of the morphology recognition model.

[0008] Specifically, the antenna array consists of multiple antenna elements; one of the antenna elements is installed at the bottom or side of each storage compartment; The RFID precision identification subsystem also includes a radio frequency switch matrix, which is connected between the antenna array and the multi-channel reader / writer. The multi-channel reader sends control commands to the radio frequency switch matrix. After receiving the control commands, the radio frequency switch matrix turns on each antenna unit in turn according to a set time period to obtain the received signal strength indication value of the radio frequency tag in the storage compartment where each antenna unit is located. The intelligent storage management platform compares the received signal strength indicator value with the set strength benchmark value. When the received signal strength indicator value is higher than the set strength benchmark value, it determines that the confidential carrier is stored in the storage compartment.

[0009] Specifically, the 3D-printed custom storage structure includes a main cavity and a mounting base; The main cavity is constructed using a honeycomb-shaped hollow support structure, and the inner surface of the main cavity is provided with an inwardly protruding contact portion. The mounting base is located at the bottom or side of the main cavity, and the mounting base is equipped with standard mechanical connectors. The 3D printed custom storage structure frame is fixed to the internal shelf or inner wall of the cabinet through the standard mechanical connectors.

[0010] Specifically, the intelligent storage management platform also includes an anomaly monitoring and alarm module; The intelligent matching module sends the location coordinates of the target storage compartment to the anomaly monitoring and alarm module. The anomaly monitoring and alarm module obtains the actual location coordinates of the storage compartments whose status has changed, which are uploaded in real time by the multi-channel reader. When the actual location coordinates are inconsistent with the location coordinates of the target storage compartment, the abnormal monitoring alarm module generates a misplacement alarm signal, triggering the audible and visual alarm device to issue an audible and visual alarm prompt.

[0011] Specifically, each antenna element of the antenna array is integrally provided with a parasitic resonant element on its side; The parasitic resonant unit is mechanically coupled to the bottom or side of the storage compartment; The multi-channel reader periodically performs broadband impedance matching frequency scanning to obtain the current optimal resonant frequency point of each antenna element, and sends the optimal resonant frequency point to the intelligent storage management platform.

[0012] Specifically, the intelligent storage management platform includes a deformation monitoring module; The deformation monitoring module extracts the optimal resonant frequency value of the same antenna element within multiple consecutive cycles. The deformation monitoring module calculates the frequency drift gradient of the optimal resonant frequency point. When the frequency drift gradient exceeds the set creep warning threshold, the deformation monitoring module determines that the storage compartment where the antenna unit is located has undergone physical deformation.

[0013] Specifically, the intelligent storage management platform also includes an adaptive compensation module; When the deformation monitoring module determines that the storage compartment has undergone physical deformation, the adaptive compensation module adjusts the operating frequency of the corresponding antenna unit to the current optimal resonant frequency. The adaptive compensation module increases the radio frequency transmission power of the antenna unit.

[0014] Specifically, the intelligent storage management platform includes a hot spot early warning module; When the intelligent storage management platform determines that the confidential carrier is stored in the storage compartment, the hot spot early warning module sends a pilot signal to the multi-channel reader. The multi-channel reader injects radio frequency pilot signals into the corresponding antenna unit through the radio frequency switch matrix; The multi-channel reader measures and uploads the impedance mismatch parameters of the antenna port of the antenna unit to the hot spot early warning module.

[0015] Specifically, the hot spot early warning module calculates the sudden increase gradient of the impedance mismatch parameter; When the sudden increase in gradient exceeds the set reference impedance change threshold, the hot spot early warning module determines that the corresponding storage compartment has experienced transient mismatch. The hot spot early warning module will delay the action of sending control commands to the antenna unit until after a preset thermal attenuation time, or instruct the multi-channel reader to switch to a backup operating frequency to obtain the identification information and the received signal strength indication value.

[0016] Specifically, the RFID precise identification subsystem operates in either a low-frequency band or a high-frequency band; when operating in a low-frequency band, the low-frequency band is 120kHz to 134.2kHz; when operating in a high-frequency band, the operating frequency is 13.56MHz; the RFID tag is a passive tag; the antenna array includes multiple coil antenna units, and the multi-channel reader transmits energy and data to the RFID tag through the coil antenna units via inductive coupling. Further, the RFID precise identification subsystem operates in a low-frequency band, specifically at a frequency of 125kHz.

[0017] The advantages of this invention compared to existing technologies lie in its deep integration of underlying physical structure and upper-level intelligent algorithms, breaking the limitations of traditional fixed compartments. By setting up and applying a 3D-printed customized storage structure frame within the cabinet, coupled with an intelligent matching architecture composed of image acquisition equipment and shape recognition models, the system can automatically acquire the size and shape data of the confidential carrier to be stored, and accurately output the target storage compartment using a built-in genetic algorithm solver. Combined with the linkage between the multi-channel reader / writer and the radio frequency switch matrix in the RFID precision identification subsystem, the system employs a technical feature of sequentially activating each antenna unit according to a set time period, eliminating signal crosstalk caused by multiple items densely stored in the same storage compartment. This achieves a leap from one item per compartment to precise positioning of multiple items per compartment, significantly improving space utilization and error prevention management capabilities.

[0018] To address the problem of slow aging and deformation of physical structures, this invention utilizes the electromagnetic principle of antenna resonant frequency drift caused by dielectric deformation. By integrating a parasitic resonant unit next to the antenna element and mechanically coupling it with the storage compartment, the multi-channel reader can periodically perform broadband impedance matching frequency scanning to obtain the optimal resonant frequency point. When the frequency drift gradient calculated by the deformation monitoring module exceeds the set creep warning threshold, the system adaptively compensates for the RF offset caused by the micromechanical creep of the 3D printed material by adjusting the operating frequency to the current optimal resonant frequency point and increasing the RF transmission power. This eliminates the maintenance blind spot of misjudging structural physical deformation as hardware loss of electronic components.

[0019] To address the transient RF dead zone phenomenon caused by the secure carrier with active components, this invention introduces the RF physical law of impedance transient mismatch caused by local hot spot effect. When the system detects the carrier's presence, it rapidly calculates the sudden increase gradient to identify the transient mismatch by injecting RF pilot signals into the corresponding antenna element and measuring the impedance mismatch parameters at the antenna port. Subsequently, the system intelligently delays the transmission of control commands until a preset thermal attenuation time, or switches the commands to a backup operating frequency for identification. This completely eliminates the transient read-miss phenomenon caused by thermal stress without requiring complex physical heat dissipation hardware. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the overall structure and module relationship of the intelligent storage system for confidential carriers that integrates RFID and 3D printing, showing the composition and connection relationship of the cabinet, 3D printing rack, imaging equipment and RFID reader / writer, and intelligent management platform.

[0021] Figure 2 This is a flowchart of the training process of the morphology recognition model of the present invention, which shows the closed-loop process from dataset acquisition and forward propagation of the convolutional neural network to loss function calculation and backpropagation to update weights.

[0022] Figure 3 This is a cross-sectional schematic diagram of the actual structure of the 3D printed customized storage structure frame of the present invention, showing the main cavity, honeycomb hollow support structure, inwardly protruding contact part, and mounting base for fixing.

[0023] Figure 4 This is a partial view of the storage compartment for the confidential carrier of this invention and the distribution of RFID hardware, showing the connection between the carrier, the radio frequency tag and the antenna unit, radio frequency switch matrix and multi-channel reader at the bottom of the compartment.

[0024] Figure 5 This is a scene mechanism diagram of the abnormal monitoring and alarm of the present invention. By comparing the target location with the actual location coordinates, an audible and visual alarm device is triggered to provide a prompt when a discrepancy occurs.

[0025] Figure 6 This is a schematic diagram of the mechanical coupling and deformation monitoring of the parasitic resonant unit of the present invention, which illustrates how the bottom deformation affects the antenna unit and the parasitic resonant unit through mechanical coupling, thereby causing the resonant frequency to drift beyond the threshold.

[0026] Figure 7 This is a network topology diagram of the RFID antenna polling and conduction of the present invention, which shows the implementation method of the multi-channel reader periodically and alternately conducting multiple antenna elements through an RF switch matrix. Detailed Implementation

[0027] The present invention will now be described with reference to the accompanying drawings. The following content is intended to help those skilled in the art understand and implement the present invention, and is not intended to limit the specific scope of protection of the present invention. Without departing from the inventive concept, the cabinet dimensions, number of compartments, antenna arrangement, platform deployment method, and software parameters can all be adjusted according to the actual usage scenario.

[0028] In one embodiment, the present invention provides an intelligent storage system for secure carriers that integrates RFID and 3D printing, such as... Figure 1 As shown, the system includes a cabinet, image acquisition equipment, an RFID precision identification subsystem, an intelligent storage management platform, and a 3D-printed custom storage structure frame installed inside the cabinet.

[0029] The cabinet serves as the physical storage space for confidential materials. It can be made of metal or composite materials. For applications requiring high levels of security, an electromagnetic shielding layer can be added. The cabinet's interior features shelves or mounting rails to secure the 3D-printed custom storage structure. Power and communication interfaces can also be pre-installed within the cabinet to ensure the stable operation of image acquisition equipment, RFID identification components, and the intelligent storage management platform.

[0030] The image acquisition device is connected to the intelligent storage management platform. The image acquisition device can be installed inside the cabinet door, at the cabinet entrance, or on a separate storage registration desk. Before storing the secure container, the operator places it in the image acquisition area. The image acquisition device captures an image of the container's appearance and sends the image data to the intelligent storage management platform. The image acquisition device can be an industrial camera or a depth camera. For scenarios requiring thickness information, a binocular camera or a camera module with structured light ranging capability is preferred.

[0031] The 3D-printed custom storage structure has multiple independent storage compartments. Each compartment is used to hold confidential materials. It should be noted that the cabinet itself does not have 3D printing capabilities; the 3D-printed custom storage structure (i.e., the storage compartments for confidential materials) is custom-printed by an external, independent 3D printing device according to the different shapes of the confidential materials, and then installed internally into the cabinet. Under normal circumstances, one compartment holds one confidential material; however, after the platform completes size matching and RFID verification, one compartment can also hold multiple confidential materials permitted by the platform. The purpose of using a 3D-printed custom structure is to overcome the problem of the limited size of traditional fixed compartments. Traditional cabinets typically use uniform molds to form compartments, which results in small-sized materials occupying too much space, while irregularly shaped materials are difficult to fit. This invention uses 3D printing to create custom compartments, allowing for personalized design based on the size distribution, shape type, and internal space of the confidential materials, ensuring that confidential materials of different shapes receive suitable support and positioning space.

[0032] 3D printing materials can include nylon, carbon fiber reinforced nylon, ABS, PC, or PEEK. For classified cabinets that operate for extended periods, engineering plastics with good flame retardant properties and relatively stable dielectric properties are preferable. This ensures structural strength while minimizing the impact of material aging on RFID performance.

[0033] The RFID precision identification subsystem includes radio frequency tags (RFID tags), an antenna array, and a multi-channel reader. RFID tags are mounted on RFID clips, which are connected to the secure carrier via retractable cords. Each RFID clip corresponds to a dedicated slot within the cabinet, located in a specific storage compartment. The antenna unit identifies the RFID tag on the clip at the corresponding slot. Since each slot corresponds one-to-one with a storage compartment, the RFID tag's identification result can be mapped to the corresponding compartment. Unique identification information is written into or bound to the RFID tag. This information can correspond to the carrier number or be associated with business data such as security level, responsible person, registration time, carrier type, and circulation status. Passive RFID tags are preferred, and their operating frequency band matches that of the multi-channel reader: in the preferred low-frequency embodiment, low-frequency passive tags are used; in the high-frequency embodiment, high-frequency passive tags are used; and in the ultra-high-frequency embodiment, ultra-high-frequency passive tags are used. For metal-encapsulated secure carriers, anti-metal tags can be used. For paper document bags, flexible tags can be used.

[0034] The RFID identification in this invention can be implemented using various operating frequency bands, such as low frequency, high frequency, or ultra-high frequency. In other words, this invention includes multiple embodiments employing different RFID operating frequency bands. In the preferred low-frequency embodiment, the RFID precision identification subsystem operates in the low-frequency (LF) band, typically between 120kHz and 134.2kHz, with 125kHz being a typical example. Low-frequency RFID operates based on the principle of inductive coupling. The antenna unit on the reader side and the tag coil within the RFID tag are coupled through a near-field alternating magnetic field to transmit energy and data. Low-frequency signals are relatively less affected by metal and liquid environments, and the near-field magnetic coupling has a short range and clear boundaries, naturally suppressing cross-reading between adjacent compartments. This makes it particularly suitable for scenarios involving complex metal environments within cabinets, short identification distances, and requiring precise compartment-level positioning for secure carrier storage. In high-frequency embodiments, the RFID precision identification subsystem operates in the high-frequency (HF) band, typically at 13.56MHz. It also operates based on the principle of inductive coupling, and the tag protocol standard is mature (e.g., ISO / IEC 15693). It has strong multi-tag anti-collision capabilities and is suitable for scenarios with high requirements for reading speed and batch inventory capabilities. In ultra-high-frequency embodiments, the RFID precision identification subsystem operates in the ultra-high-frequency (UHF) band, such as the 860MHz to 960MHz range permitted by local regulations. It operates based on the principle of electromagnetic backscattering, resulting in a longer identification distance. In this case, the antenna unit can be a microstrip antenna. The specific operating frequency band used is determined based on the type of security carrier, the electromagnetic environment of the cabinet, the identification distance requirements, and the spectrum regulations of the region of use. Unless otherwise specified, the subsequent descriptions of antenna types, scanning bands, and frequency parameters in this specification are adapted to the selected operating frequency band.

[0035] The antenna array is distributed within the storage compartments of the 3D-printed custom storage structure. A multi-channel reader is connected to the antenna array. The multi-channel reader collects the identification information of the RFID tags and simultaneously acquires the received signal strength indicator value, sending the results to the intelligent storage management platform. The received signal strength indicator value is used to determine the distance relationship between the RFID tag and the corresponding compartment antenna and the reliability of the reading. By reading the identification information, the platform can identify the corresponding secure carrier. By reading the received signal strength indicator value, the platform can further determine whether the secure carrier is located within the corresponding storage compartment.

[0036] The intelligent storage management platform comprises a digital twin mapping module and an intelligent matching module. The digital twin mapping module is used to construct a 3D visualization model of the 3D-printed customized storage structure. This 3D visualization model can be generated from CAD design files or from the imported modeling files of the 3D-printed structure. The platform assigns a unique compartment number to each storage location and establishes a correspondence between the compartment number, its location within the cabinet, its 3D coordinates, available dimensions, and antenna channel number. After the secure carrier is stored, the digital twin mapping module binds the RFID tag's identification information to the corresponding storage compartment in the 3D visualization model. Managers can intuitively see the presence and spatial location of each secure carrier on the platform interface.

[0037] The intelligent matching module selects a target storage compartment for the confidential carrier to be stored. Once the carrier enters the image acquisition area, the image acquisition device captures its image data. The intelligent matching module inputs the image data into the shape recognition model. The shape recognition model outputs the size and shape data of the confidential carrier to be stored. This data may include length, width, and thickness, as well as outer contour type, edge protrusion status, surface material characteristics, whether it has a metal casing, and whether there are external cables or protruding interfaces.

[0038] For ordinary paper documents, the shape recognition model primarily outputs the dimensions and thickness of the circumscribed rectangle. For external hard drives or encrypted cases, the shape recognition model can also identify the casing material and the area of ​​maximum thickness. For irregularly shaped electronic media, the shape recognition model can further output the supportable area and the non-compressible area. The purpose of this design is to ensure that subsequent compartment matching not only determines whether the compartment can be placed, but also whether it is stable after placement, whether it will compress the antenna, and whether it will affect the RFID of adjacent compartments.

[0039] After the morphology recognition model outputs size and shape data, the intelligent matching module inputs this data into the built-in genetic algorithm solver. The genetic algorithm solver calculates and outputs the target storage space based on the current availability of storage spaces, space size, storage capacity, reliability of historical data reads, occupancy status of adjacent storage spaces, and carrier access frequency.

[0040] The reason for employing a genetic algorithm solver is that matching the storage compartments for secure carriers is not a simple size comparison problem. Selecting only the compartments that fit perfectly might result in larger compartments being occupied by smaller carriers, reducing subsequent space utilization. Conversely, selecting only the closest compartment might lead to multiple highly reflective tags clustering in the same area, increasing the risk of cross-reading and missed reads. The genetic algorithm can perform global optimization among multiple candidate compartments, achieving a balance between space utilization, RFID reliability, and ease of retrieval.

[0041] In one embodiment, the genetic algorithm solver represents a candidate solution as the storage location number corresponding to the confidential carrier to be stored. In batch storage scenarios, a candidate solution can also be represented as an allocation sequence between multiple confidential carriers and multiple storage locations. The initial population can be generated by random storage location allocation, size-first allocation, or by combining storage locations with high historical read reliability. During fitness evaluation, the platform comprehensively calculates the remaining space in the storage location, the carrier size matching degree, the distance between the tag and the antenna element, the number of carriers already stored in adjacent storage locations, and the number of historical misreads of the storage location. For heavier carriers, a storage location carrying capacity level can also be introduced. The population size can be 20 to 200. The number of iterations can be 30 to 300. The crossover probability can be 0.5 to 0.9. The mutation probability can be 0.01 to 0.2. When the fitness improvement is less than a preset range for 5 to 30 consecutive rounds, or when the maximum number of iterations is reached, the genetic algorithm solver outputs the target storage location.

[0042] In a further embodiment, such as Figure 2 As shown, the training process of the morphological recognition model includes sample acquisition, manual annotation, forward propagation, loss calculation, and backpropagation update. First, a set of sample images of secure carriers is acquired, and corresponding actual size and morphological labels are manually labeled. Sample images can include images of paper document bags, portable hard drives, USB flash drives, CD cases, magnetic tape cases, encrypted terminals, and metal-packaged carriers. To improve the model's adaptability, sample images can be acquired under different lighting conditions, as well as from different angles and against different backgrounds.

[0043] Manually labeled actual size and shape attributes include length, width, and thickness. For irregularly shaped carriers, these attributes may also include outer contour category, placement orientation, and protruding part information. After the training set is constructed, it is fed into the initial deep convolutional neural network. The deep convolutional neural network extracts features from the image and calculates predicted size and shape data through forward propagation.

[0044] Deep convolutional neural networks can employ either the ResNet or MobileNet architecture. For situations where the computing power of the internal cabinet is limited, a lightweight convolutional neural network is preferred. The network front-end extracts edge and texture features. The middle layers extract overall contour features. The network back-end can be configured with regression and classification branches. The regression branch outputs length, width, and thickness. The classification branch outputs the carrier's outer contour type and material type.

[0045] The platform calculates the loss function value between the predicted size and shape data and the actual size and shape labels. The loss function can be formed by combining size regression loss and shape classification loss. The size regression loss can use mean absolute error or mean squared error. The shape classification loss can use cross-entropy loss. Then, the backpropagation algorithm is applied to calculate the gradient of the loss function value with respect to the weights of each layer node in the deep convolutional neural network, and the weights of each layer node are updated. The optimization algorithm can use stochastic gradient descent or the Adam algorithm. The learning rate can be between 0.00001 and 0.01. The batch size can be between 8 and 128. The number of training epochs can be between 50 and 500. When the loss function value is less than the set convergence threshold, the training of the shape recognition model is complete. The convergence threshold can be between 0.001 and 0.05. Through this training process, the shape recognition model can identify the size and shape of the confidential carrier from the image, providing reliable input for the intelligent matching module.

[0046] In one embodiment, to further improve size recognition accuracy, a fixed size reference object or calibration plate can be set in the image acquisition area. The platform performs scale correction on the model output based on the proportional relationship between the pixel size of the calibration plate in the image and the actual size. For metal carriers with strong surface reflectivity, diffuse reflection supplementary lighting can be set in the image acquisition area. This can reduce the impact of highlight areas on contour recognition.

[0047] In a further embodiment, the antenna array consists of multiple antenna elements. One antenna element is located at the bottom or side of each storage compartment, such as... Figure 4 As shown. In the preferred low-frequency and high-frequency embodiments, the antenna unit is a coil antenna unit. The coil antenna unit can be a wound coil or a planar spiral coil etched or printed on a circuit board, and can be embedded in a reserved slot of the 3D-printed customized storage structure frame. In the ultra-high frequency embodiment, the antenna unit can be a microstrip antenna unit, such as a rectangular patch antenna, a circular patch antenna, an inverted F structure antenna, or a slot-coupled antenna. The antenna unit is preferably located close to the dedicated card slot, so that the coupled magnetic field or radio frequency energy is concentrated to cover the card slot area of ​​the corresponding storage compartment, thereby completing the identification of the RFID tag on the RFID card at the card slot and mapping the identification result to the corresponding storage compartment.

[0048] The RFID precision identification subsystem also includes an RF switch matrix. The RF switch matrix is ​​connected between the antenna array and the multi-channel reader, such as... Figure 7As shown. In the low-frequency and high-frequency embodiments, the RF switch matrix can be implemented using an analog multiplexer or a relay array; in the ultra-high-frequency embodiment, the RF switch matrix can be implemented using an RF switch chip. The multi-channel reader sends control commands to the RF switch matrix. After receiving the control commands, the RF switch matrix turns on each antenna element in turn according to a set time period. Each time an antenna element is turned on, the multi-channel reader reads the identification information of the RFID tag in the corresponding storage compartment and obtains the received signal strength indication value.

[0049] The advantage of alternating conduction is that it avoids multiple storage unit antennas operating simultaneously. If multiple antennas operate concurrently, electromagnetic crosstalk can easily occur between adjacent storage units, potentially causing the platform to read tags from locations other than the target storage unit. This invention uses an RF switch matrix to sequentially activate antenna units, ensuring that each identification corresponds to a specific storage unit, thereby improving storage unit-level positioning accuracy. The activation time of a single antenna unit can range from 10 milliseconds to 500 milliseconds. The overall cabinet polling cycle can range from 0.5 seconds to 30 seconds. For frequent access scenarios, the polling cycle can be shortened. For long-term static monitoring scenarios, the polling cycle can be extended to reduce power consumption and RF heat generation.

[0050] The intelligent storage management platform compares the received signal strength indication value with a set strength benchmark value. When the received signal strength indication value is higher than the set strength benchmark value, the platform determines that the secure carrier should be stored in that storage compartment. Here, "higher than" means that the radio frequency signal strength has reached a stable and identifiable range; for example, in dBm, -45dBm is higher than -65dBm. In low-frequency and high-frequency embodiments, the received signal strength indication value can be characterized by the amplitude of the induced signal or the demodulated signal detected by the reader's receiving front end. The strength benchmark value is set accordingly using the amplitude value or its logarithm, and the comparison method is the same. The strength benchmark value can be obtained through factory calibration or by the platform's self-learning after on-site installation. Typically, the strength benchmark value can be set to be 5dB to 20dB higher than the ambient noise floor. It can also be set in the range of -75dBm to -35dBm depending on the specific tag type. For metal-cased carriers, the strength benchmark value can be appropriately reduced. For areas with a high risk of cross-reading between adjacent compartments, the strength benchmark value can be appropriately increased.

[0051] To avoid misjudgments caused by fluctuations in a single read, the intelligent storage management platform can employ a multi-period confirmation mechanism. For example, if the same RFID tag is read for 2 to 5 consecutive polling cycles, and the received signal strength indicator value is consistently higher than the strength benchmark value, the platform confirms that the secure carrier has been stored in the compartment. If the RFID tag is not read for 2 to 5 consecutive polling cycles, or the received signal strength indicator value remains below the strength benchmark value, the platform confirms that the secure carrier has been removed. In this way, the brief read signals generated when operators carry the carrier through the cabinet area can be filtered out.

[0052] In a further embodiment, the 3D-printed custom storage structure includes a main cavity and a mounting base, such as... Figure 3 As shown, the main cavity is constructed using a honeycomb-shaped perforated support structure. This structure reduces material usage while maintaining load-bearing strength, and also lowers the weight of the structural frame. The honeycomb pores also provide channels for air circulation within the cabinet, reducing localized heat buildup. The honeycomb pores can be hexagonal or rounded polygonal. The pore diameter can range from 3 mm to 30 mm. The wall thickness can range from 1 mm to 6 mm. For heavier metal encapsulation carriers, the honeycomb structure can be locally densified, and reinforcing ribs can be added to stress-bearing areas.

[0053] The inner surface of the main cavity is provided with inwardly protruding contact parts. These contact parts can be made into dot-shaped bosses or strip-shaped support ribs. For carriers with relatively regular shapes, the contact parts can be located in the fixed support areas at the bottom and side walls of the compartment. For irregularly shaped carriers, the contact parts can be customized as curved support parts according to the carrier's outer contour. The purpose of setting the contact parts is to provide stable support for the secure carrier after it is placed in the compartment and to prevent the carrier from directly pressing on the antenna unit or parasitic resonant unit. The contact parts also reduce the large area of ​​contact between the carrier surface and the inner wall of the compartment, making it easier for operators to handle.

[0054] The mounting base is located at the bottom or side of the main cavity. The mounting base is equipped with standard mechanical connectors. These connectors can be screw holes or snap-fit ​​fasteners. Depending on the cabinet structure, guide rails, locating pin holes, or quick-release latches can also be used. The 3D-printed custom storage rack is fixed to the internal shelves or inner wall of the cabinet using these standard mechanical connectors. The purpose of using standard mechanical connectors is to improve the replaceability of the rack. If the type of security device changes in the future, only a new storage rack needs to be redesigned and printed, while retaining the original cabinet mounting interface; the entire cabinet does not need to be replaced.

[0055] In one embodiment, the 3D-printed customized storage structure frame is designed using 3D modeling software before printing. During modeling, the storage compartments, antenna unit mounting positions, cable trays, parasitic resonant unit mounting surfaces, contact parts, and compartment numbering are all incorporated into the structural model. After printing, the structure frame undergoes support removal and surface cleaning. Areas requiring high precision, such as antenna unit mounting positions, can be locally machined. The antenna feed cable is laid along the pre-reserved cable tray and secured with a cable clamping structure to prevent cable movement during long-term use and thus maintain RF impedance stability.

[0056] In a further embodiment, the intelligent storage management platform also includes an anomaly monitoring and alarm module, such as... Figure 5As shown. After the intelligent matching module outputs the target storage location, it sends the location coordinates of the target storage location to the anomaly monitoring and alarm module. The location coordinates can be represented using three-dimensional coordinates, or a combination of cabinet number, floor number, column number, and storage location number. During the actual storage process, the multi-channel reader uploads the actual location coordinates of storage locations whose status has changed in real time. The anomaly monitoring and alarm module compares the actual location coordinates with the location coordinates of the target storage location.

[0057] When the actual location coordinates differ from the target storage compartment's location coordinates, the anomaly monitoring and alarm module generates a misplacement alarm signal and triggers an audible and visual alarm device. The audible and visual alarm device can include a buzzer and indicator lights, or it can include an LED light strip inside the cabinet and a pop-up window on the screen. Alarm records can also be simultaneously written to the platform log, which records the confidential carrier identifier, operator identity, target compartment, actual compartment location, and time of occurrence. To avoid false alarms triggered when the carrier briefly passes through other compartments during movement, the anomaly monitoring and alarm module can be configured with a confirmation delay. The confirmation delay can range from 1 to 10 seconds. The platform will only trigger a misplacement alarm if the actual location coordinates remain inconsistent with the target location coordinates for more than the confirmation delay.

[0058] In a further embodiment, a parasitic resonant unit is integrally disposed beside each antenna element of the antenna array. In preferred low-frequency and high-frequency embodiments, the parasitic resonant unit can be a parasitic resonant ring composed of a resonant coil and a resonant capacitor; in ultra-high-frequency embodiments, the parasitic resonant unit can be a parasitic patch. The parasitic resonant unit can be disposed on the same dielectric substrate as the antenna element, or it can be embedded in a reserved slot in the 3D printed structural frame. The parasitic resonant unit is mechanically coupled to the bottom or side of the storage compartment, such as... Figure 6 As shown. Here, mechanical coupling refers to the fact that when the bottom or side of the storage compartment bends, is compressed, warps, or slowly creeps, the parasitic resonant unit also changes its relative position due to structural changes. Changes in the distance between the parasitic resonant unit and the antenna unit affect the electromagnetic coupling environment around the antenna unit, thus causing a drift in the optimal resonant frequency.

[0059] The multi-channel reader periodically performs broadband impedance matching frequency scanning to obtain the current optimal resonant frequency point of each antenna element and sends this optimal resonant frequency point to the intelligent storage management platform. Broadband impedance matching frequency scanning can be achieved through the reader's built-in RF front-end or through a directional coupler and reflected power detection circuit. The scanning frequency band is determined based on the RFID operating frequency band used: in the preferred low-frequency embodiment, the scanning frequency band can cover 110kHz to 140kHz, with a scanning step of 0.05kHz to 1kHz; in the high-frequency embodiment, the scanning frequency band can cover 12MHz to 15MHz, with a scanning step of 1kHz to 50kHz; in the ultra-high-frequency embodiment, the scanning frequency band can cover 840MHz to 960MHz, determined according to the permitted RFID operating frequency band in the area of ​​use, with a scanning step of 50kHz to 2MHz. The scanning cycle can be 1 hour to 24 hours. For newly installed 3D printed structural frames, the system can increase the scanning frequency for the first 7 to 30 days to observe frequency changes caused by initial stress release in the material.

[0060] The intelligent storage management platform includes a deformation monitoring module. This module extracts the optimal resonant frequency value of the same antenna element over multiple consecutive periods and calculates the frequency drift gradient of that optimal resonant frequency value. The frequency drift gradient represents how quickly the optimal resonant frequency changes over time. The number of consecutive periods can be 3 to 20. The calculation time window can be 6 hours to 30 days. The creep warning threshold can be set to 0.01% to 2% relative to the initial resonant frequency; when set in absolute value, it can be 0.01kHz / day to 2kHz / day in the preferred low-frequency embodiment, 1kHz / day to 200kHz / day in the high-frequency embodiment, and 0.1MHz / day to 10MHz / day in the ultra-high-frequency embodiment. When the frequency drift gradient exceeds the set creep warning threshold, the deformation monitoring module determines that the storage compartment containing the antenna element has undergone physical deformation.

[0061] This assessment method leverages the sensitivity of the antenna's resonant frequency to changes in the surrounding dielectric environment and geometric position. Traditional devices often only check the reader or tag after a read failure, easily overlooking radio frequency offset caused by slow deformation of the supporting structure. This invention, through parasitic resonant units and frequency drift monitoring, can identify potential risks before structural deformation is obvious.

[0062] In a further embodiment, the intelligent storage management platform also includes an adaptive compensation module. When the deformation monitoring module determines that a storage compartment has undergone physical deformation, the adaptive compensation module adjusts the operating frequency of the corresponding antenna unit to a frequency close to the current optimal resonant frequency. For multi-channel readers that support frequency point configuration, the adaptive compensation module can issue frequency point adjustment commands, enabling the reader to preferentially select an operating frequency close to the optimal resonant frequency within the limits permitted by regulations. In ultra-high frequency embodiments, for readers employing frequency hopping mechanisms, the residing ratio of frequency bands near the optimal resonant frequency can be increased.

[0063] The adaptive compensation module also increases the RF transmit power of the corresponding antenna element. The increase in RF transmit power can range from 1dB to 6dB, or it can be increased in increments of 5% to 50% of the original transmit power. Power adjustments should not be too large at once to avoid crosstalk between adjacent compartments, and the adjusted RF transmit power should not exceed the equipment certification range and the limits permitted by regulations in the region of use. A preferred approach is to increase the power in stages, 1dB to 2dB each time, and to re-check the received signal strength indicator value after each adjustment.

[0064] In a further embodiment, the intelligent storage management platform includes a hot spot early warning module. When the intelligent storage management platform determines that a secure carrier is stored in a storage compartment, the hot spot early warning module sends a pilot command to the multi-channel reader. The multi-channel reader injects a radio frequency pilot signal into the corresponding antenna unit through a radio frequency switch matrix. The pilot signal can be a low-power continuous wave or a short-pulse radio frequency signal. The purpose of this pilot signal is not to read the RFID tag, but to detect the current port impedance state of the antenna unit. The multi-channel reader measures the impedance mismatch parameters of the antenna unit's antenna port and uploads the measurement results to the hot spot early warning module. The impedance mismatch state of the antenna port can be characterized by impedance mismatch parameters, which can be voltage standing wave ratio, reflection coefficient modulus, or return loss; in preferred low-frequency and high-frequency embodiments, the impedance mismatch parameters can also be characterized by the resonant frequency detuning of the resonant circuit, the change in quality factor, or the change in the amplitude of the induced signal.

[0065] The reason for setting up a hot spot early warning module is that some secure carriers have independent power supplies, processors, or high-speed storage components. When these carriers are transferred from the working state to the cabinet, residual heat may remain on their surface. The internal space of the cabinet is relatively enclosed, making it difficult for heat to dissipate quickly. Localized temperature increases can change the dielectric properties of materials near the antenna and alter the reflection state of radio frequency signals by the metal casing, thus causing transient impedance mismatch at the antenna port. If high-frequency readings are performed immediately at this time, the system may experience brief missed readings and mistakenly identify the tag as damaged or the carrier as displaced. This invention can identify this transient mismatch state in advance by measuring pilot signals and impedance mismatch parameters.

[0066] In a further embodiment, the hot spot early warning module calculates the sudden increase gradient of the impedance mismatch parameter. The sudden increase gradient represents the rate at which the impedance mismatch parameter rises within a short period of time. The platform can perform multiple samplings within the first 10 to 300 seconds after the carrier is stored. The sampling interval can be 0.5 to 10 seconds. When the voltage standing wave ratio (VSWR) characterizes the impedance mismatch parameter, the reference impedance change threshold can be set to an increase of 0.2 to 2.0 in the VSWR per unit sampling period. Alternatively, it can be set to an increase of 10% to 100% relative to the reference value before storage. When the sudden increase gradient exceeds the set reference impedance change threshold, the hot spot early warning module determines that a transient mismatch has occurred in the corresponding storage compartment.

[0067] When a transient mismatch occurs, the hot spot warning module delays sending control commands to the antenna unit until after a preset thermal attenuation time. The preset thermal attenuation time can range from 30 seconds to 1800 seconds. For external hard drives or encrypted terminals that generate significant heat, a larger value can be used. For ordinary paper document bags, a smaller value can be used. Alternatively, the hot spot warning module instructs the multi-channel reader to switch to a backup operating frequency to continue acquiring the RFID tag's identification information and received signal strength indication. The backup operating frequency is within the allowable range of the adopted RFID operating frequency band and avoids frequencies with significant impedance mismatch. In UHF embodiments, the backup operating frequency can be a frequency in a backup frequency hopping band. Through delay control or backup operating frequency identification, the system does not require additional complex heat dissipation hardware and can reduce transient missed reads caused by thermal stress.

[0068] A complete storage process is as follows: After completing identity authentication, the operator places the secure carrier to be stored in the image acquisition area. The image acquisition device acquires the image data of the secure carrier and sends it to the intelligent storage management platform. The shape recognition model identifies the size and shape data of the carrier. The intelligent matching module runs a genetic algorithm solver based on the size and shape data and the current storage space occupancy status, outputting the target storage space. The digital twin mapping module displays the target storage space in the 3D visualization model. The operator places the secure carrier into the target storage space according to the platform prompts. The multi-channel reader activates the corresponding antenna unit through the RF switch matrix, reading the RFID tag's identification information and the received signal strength indication value. If the received signal strength indication value is higher than the strength benchmark value, the platform confirms that the secure carrier is in place and binds the RFID tag's identification information to the corresponding storage space in the 3D visualization model.

[0069] If an operator places a secure carrier in a non-target storage location, the anomaly monitoring and alarm module will receive the actual location coordinates of the changed status. If these actual location coordinates do not match the location coordinates of the target storage location, the platform will trigger a misplacement alarm. After the operator corrects the storage location, the platform will reread the RFID tag and update the digital twin mapping relationship. Through this process, the registration of secure carriers entering the cabinet, storage location selection, actual storage, and misplacement correction can form a closed loop.

[0070] A complete long-term monitoring process is as follows: The system reads RFID tags one by one according to a set polling cycle and continuously updates the location status of each secure carrier. A multi-channel reader performs broadband impedance matching frequency scanning according to a set scanning cycle to obtain the optimal resonant frequency point for each antenna element. The deformation monitoring module performs trend analysis on the optimal resonant frequency points over multiple consecutive cycles. When the frequency drift gradient exceeds the creep warning threshold, the platform determines that physical deformation has occurred in the corresponding storage location. The adaptive compensation module then adjusts the operating frequency point of the antenna element in its operating band to be close to the current optimal resonant frequency point and appropriately increases the RF transmission power. This process enables the system to detect latent deformations before structural damage expands and maintains identification stability through RF compensation.

[0071] A complete hotspot handling process is as follows: Once the platform confirms that a secure carrier has just been stored in a storage compartment, the hotspot warning module controls the multi-channel reader to inject pilot signals into the corresponding antenna unit. The multi-channel reader measures the impedance mismatch parameters at the antenna port and uploads the measurement results to the hotspot warning module. The hotspot warning module calculates the sudden increase gradient of the impedance mismatch parameters. If the sudden increase gradient exceeds the reference impedance change threshold, the platform determines that there is a transient mismatch in the compartment. At this time, the platform can wait for a preset thermal attenuation time before executing the control command. The platform can also switch to a backup operating frequency to continue identifying RFID tags. This avoids short-term RF dead zones caused by residual heat from newly placed active secure carriers, thereby reducing false alarms and missed reads.

[0072] In a further embodiment, the digital twin mapping module can also display the health status of storage locations. Storage locations without carriers are displayed as idle. Storage locations with carriers are displayed as in-place. Storage locations where carriers have been misplaced are displayed as abnormal. Storage locations where frequency drift exceeds limits are displayed as maintenance warning status. Storage locations where transient mismatch occurs are displayed as heat decay waiting status. The platform associates these statuses with operation logs and alarm logs to form a complete record of the confidential carriers from entry into storage to in-place monitoring, and then to retrieval and maintenance.

[0073] In another embodiment, multiple 3D-printed custom storage racks can be installed within a single cabinet. Different racks are adapted to different types of confidential media. For example, paper document bags use thinner and more numerous compartments. External hard drives use compartments with higher load-bearing capacity. Irregularly shaped electronic media use custom compartments with curved contact surfaces. The intelligent matching module first identifies the media type and then selects the target storage compartment from the candidate compartments of the corresponding rack. Through this partitioned matching method, the system can improve space utilization while maintaining high RFID identification reliability.

[0074] In a further embodiment, the platform can also assist in configuring RFID clips and retractable pull cords. When a new carrier is registered, the image acquisition device acquires an image of the carrier. Based on the carrier's shape, size, and material characteristics, the platform recommends a suitable RFID clip model, a retractable pull cord length, and a connection position between the pull cord and the carrier, avoiding areas prone to bending and wear. After the RFID clip is inserted into the corresponding dedicated slot, the RFID tag and the corresponding antenna unit maintain a basically fixed relative position. Identification is completed at the slot and is largely unaffected by the carrier's material or placement. By standardizing the configuration of the RFID clips and dedicated slots, the stability of the correspondence between the received signal strength indication value and the storage compartment location can be improved.

[0075] In a further embodiment, the strength reference value, creep warning threshold, reference impedance variation threshold, and thermal decay time can be preset at the factory or generated through a calibration process after on-site installation. During on-site calibration, staff place standard carriers into each storage compartment sequentially. The platform records the received signal strength indication value, optimal resonant frequency point, and impedance mismatch parameter reference value for different compartments. Based on this data, the platform generates unique reference parameters for each compartment. For equipment that has been in operation for a long time, it can be recalibrated in maintenance mode to adapt to structural aging, environmental changes, and changes in carrier type.

[0076] Through the above implementation methods, this invention integrates 3D-printed customized structures, RFID compartment-level identification, image morphology recognition, genetic algorithm matching, digital twin mapping, misplacement alarm, resonant frequency drift monitoring, adaptive RF compensation, and hot spot transient mismatch processing into a single intelligent storage system for secure carriers. This system solves the problem of traditional fixed compartments being unable to adapt to various carrier shapes and improves RFID positioning accuracy under dense storage conditions. The system can also identify RF misalignment caused by long-term mechanical creep and provide early warning and compensation for localized hot spots generated by carriers just entering the cabinet, thereby improving the reliability of secure carrier storage, identification, monitoring, and maintenance.

[0077] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A smart storage system for secure carriers integrating RFID and 3D printing, characterized in that, It includes the cabinet, image acquisition equipment, RFID precision identification subsystem, intelligent storage management platform, and 3D printed custom storage structure frame installed inside the cabinet; The image acquisition device is connected to the intelligent storage management platform; the 3D printed custom storage structure is equipped with multiple independent storage compartments, each of which is used to place a confidential carrier; The RFID precision identification subsystem includes radio frequency tags, an antenna array, and a multi-channel reader. The radio frequency tags are set on RFID clips, which are connected to a security carrier via a retractable pull cord. Each RFID clip corresponds to a dedicated slot, which is located in a corresponding storage compartment. The antenna array is distributed in the storage compartments of the 3D printed customized storage structure frame. The multi-channel reader is connected to the antenna array and collects the identification information and received signal strength indication value of the radio frequency tags and sends them to the intelligent storage management platform. The intelligent storage management platform includes a digital twin mapping module and an intelligent matching module. The digital twin mapping module constructs a three-dimensional visualization model of the 3D-printed customized storage structure and binds the identification information to the corresponding storage compartment in the three-dimensional visualization model. The intelligent matching module uses a shape recognition model to acquire image data to be stored in the secure carrier using the image acquisition device, inputs the image data into the shape recognition model, outputs the size and shape data of the secure carrier to be stored, and inputs the size and shape data into the genetic algorithm solver built into the intelligent matching module to output the target storage compartment.

2. The intelligent storage system for confidential carriers integrating RFID and 3D printing as described in claim 1, characterized in that, The training process of the morphology recognition model includes: Obtain a set of sample images of the confidential carrier and manually label them with corresponding actual size and shape tags to construct a training set; The training set is input into an initial deep convolutional neural network for feature extraction and forward propagation calculation, and the predicted size and shape data are output. Calculate the loss function value between the predicted size and shape data and the actual size and shape label; The backpropagation algorithm is used to calculate the gradient of the loss function value with respect to the weights of each layer node in the deep convolutional neural network, and the weights of each layer node are updated until the loss function value is less than the set convergence threshold, thus completing the training of the morphology recognition model.

3. The intelligent storage system for confidential carriers integrating RFID and 3D printing as described in claim 1, characterized in that, The antenna array consists of multiple antenna elements; one of the antenna elements is installed at the bottom or side of each storage compartment; The RFID precision identification subsystem also includes a radio frequency switch matrix, which is connected between the antenna array and the multi-channel reader / writer. The multi-channel reader sends control commands to the radio frequency switch matrix. After receiving the control commands, the radio frequency switch matrix turns on each antenna unit in turn according to a set time period to obtain the received signal strength indication value of the radio frequency tag in the storage compartment where each antenna unit is located. The intelligent storage management platform compares the received signal strength indicator value with the set strength benchmark value. When the received signal strength indicator value is higher than the set strength benchmark value, it determines that the confidential carrier should be stored in the storage compartment.

4. The intelligent storage system for confidential carriers integrating RFID and 3D printing as described in claim 1, characterized in that, The 3D-printed custom storage structure includes a main cavity and a mounting base; The main cavity is constructed using a honeycomb-shaped hollow support structure, and the inner surface of the main cavity is provided with an inwardly protruding contact portion. The mounting base is located at the bottom or side of the main cavity, and the mounting base is equipped with standard mechanical connectors. The 3D printed custom storage structure frame is fixed to the internal shelf or inner wall of the cabinet through the standard mechanical connectors.

5. The intelligent storage system for confidential carriers integrating RFID and 3D printing as described in claim 1, characterized in that, The intelligent storage management platform also includes an anomaly monitoring and alarm module; The intelligent matching module sends the location coordinates of the target storage compartment to the anomaly monitoring and alarm module. The anomaly monitoring and alarm module obtains the actual location coordinates of the storage compartments whose status has changed, which are uploaded in real time by the multi-channel reader. When the actual location coordinates are inconsistent with the location coordinates of the target storage compartment, the abnormal monitoring alarm module generates a misplacement alarm signal, triggering the audible and visual alarm device to issue an audible and visual alarm prompt.

6. The intelligent storage system for confidential carriers integrating RFID and 3D printing as described in claim 3, characterized in that, Each antenna element of the antenna array is integrally provided with a parasitic resonant unit on its side; The parasitic resonant unit is mechanically coupled to the bottom or side of the storage compartment; The multi-channel reader periodically performs broadband impedance matching frequency scanning to obtain the current optimal resonant frequency point of each antenna element, and sends the optimal resonant frequency point to the intelligent storage management platform.

7. The intelligent storage system for confidential carriers integrating RFID and 3D printing as described in claim 6, characterized in that, The intelligent storage management platform includes a deformation monitoring module; The deformation monitoring module extracts the optimal resonant frequency value of the same antenna element within multiple consecutive cycles. The deformation monitoring module calculates the frequency drift gradient of the optimal resonant frequency point. When the frequency drift gradient exceeds the set creep warning threshold, the deformation monitoring module determines that the storage compartment where the antenna unit is located has undergone physical deformation. The intelligent storage management platform also includes an adaptive compensation module; When the deformation monitoring module determines that the storage compartment has undergone physical deformation, the adaptive compensation module adjusts the operating frequency of the corresponding antenna unit to the current optimal resonant frequency. The adaptive compensation module increases the radio frequency transmission power of the antenna unit.

8. The intelligent storage system for confidential carriers integrating RFID and 3D printing as described in claim 3, characterized in that, The intelligent storage management platform includes a hot spot early warning module; When the intelligent storage management platform determines that the confidential carrier is stored in the storage compartment, the hot spot early warning module sends a pilot signal to the multi-channel reader. The multi-channel reader injects radio frequency pilot signals into the corresponding antenna unit through the radio frequency switch matrix; The multi-channel reader measures and uploads the impedance mismatch parameters of the antenna port of the antenna unit to the hot spot early warning module; The hot spot early warning module calculates the sudden increase gradient of the impedance mismatch parameter; When the sudden increase in gradient exceeds the set reference impedance change threshold, the hot spot early warning module determines that the corresponding storage compartment has experienced transient mismatch. The hot spot early warning module will delay the action of sending control commands to the antenna unit until a preset thermal attenuation time is reached, or instruct the multi-channel reader to switch to a backup operating frequency to obtain the identification information and the received signal strength indication value.

9. The intelligent storage system for confidential carriers integrating RFID and 3D printing as described in claim 1, characterized in that, The RFID precision identification subsystem operates in either a low-frequency band or a high-frequency band; when operating in a low-frequency band, the low-frequency band is 120kHz to 134.2kHz; when operating in a high-frequency band, the operating frequency is 13.56MHz; the RFID tag is a passive tag; the antenna array includes multiple coil antenna units, and the multi-channel reader transmits energy and data to the RFID tag via the coil antenna units in an inductive coupling manner.

10. The intelligent storage system for confidential carriers integrating RFID and 3D printing as described in claim 9, characterized in that, The RFID precision identification subsystem operates in the low-frequency band, with a working frequency of 125kHz.

Images