An automatic inventory control method and system based on mobile RFID identification

Abstract

The present application relates to the field of wireless radio frequency identification and warehouse automation logistics control technology, and particularly relates to an automatic inventory control method and system based on mobile RFID identification, comprising the following steps: S100: task configuration and height control; S200: multi-antenna time domain polling gating; S300: adaptive power radio frequency reading; S400: edge data preprocessing; S500: real-time data comparison and synchronization. When in use, the present application realizes efficient closed-loop control of automatic inventory in high-warehouse high-density warehouse environment through task-driven vertical closed-loop height adjustment, multi-antenna time domain isolation gating control, closed-loop adaptive power optimization, and mechanical and electrical and software and hardware collaborative design of edge side data cleaning pipeline. The use of closed-loop height calculation based on absolute zero point calibration and limit redundancy protection ensures the spatial alignment accuracy of the antenna assembly in the vertical axial direction, and eliminates the personal falling safety hazard of high-altitude manual inventory.

Description

Technical Field

[0001] This invention relates to the field of radio frequency identification and warehouse automation logistics control technology, specifically to an automatic inventory control method and system based on mobile RFID identification. Background Technology

[0002] Radio Frequency Identification (RFID) technology, as a non-contact automatic identification technology, uses radio frequency signals to automatically identify target objects and collect data. In recent years, it has been widely used in logistics warehousing, library archives, and asset management. Especially in large warehouse-style archives or modern high-bay racking management, in order to maximize space utilization, the vertical height of racks or mobile shelving can often reach several meters or even higher, and materials or file boxes are arranged in multiple layers, multiple columns, and extremely dense, non-hollow, compact arrangements on the racks.

[0003] However, when faced with such high-density and high-capacity special warehousing environments, existing mobile automated inventory counting devices and control systems have revealed the following core technical deficiencies in practical applications: First, the precise vertical positioning control and shelf alignment capabilities are insufficient. Faced with high-bay racks several meters high, traditional manual hand-held scanning or aerial work platform methods are not only extremely inefficient but also pose a significant risk of falls to workers. Existing lifting inventory equipment often uses open-loop control or a single incremental position sensor for its height adjustment mechanism. After a power outage and restart or long-distance reciprocating lifting and lowering, incremental sensors are prone to losing absolute position reference. Combined with the inherent slippage rate and inertial overshoot of mechanical structures such as lead screws or chain drives, this leads to cumulative errors in mechanical positioning. This prevents the antenna assembly of the inventory cart from accurately aligning with the core radiation area of ​​the target shelf shelf in the vertical axis, easily causing large-scale label misses due to height deviation, or misclassifying and mapping read labels to adjacent shelves, severely affecting the spatial accuracy of shelf inventory data.

[0004] Secondly, there are issues with concurrent interference from multiple antennas, spatial crosstalk, and false readings in high-density electromagnetic environments. In high-density mobile shelving storage scenarios, the spatial spacing between RFID tags on adjacent items is typically only 1-3 centimeters. Traditional inventory cart solutions, if employing multiple antennas to simultaneously transmit RF carriers, are prone to severe near-field electromagnetic coupling and intermodulation distortion between antennas, leading to distortion of the main lobe beam. Simultaneously, because shelves typically contain metal baffles or densely packed paper media, strong electromagnetic wave multipath reflection, diffraction, and sidelobe leakage from wide-angle antennas can cause RF energy to overflow into non-target inventory areas (such as adjacent columns or rear shelves), activating a large number of irrelevant detached tags and causing severe electromagnetic crosstalk and false readings. Simply reducing power to limit the radiation boundary can result in missed readings of deep tags located in electromagnetic blind zones or within superimposed media due to insufficient activation energy. Furthermore, the simultaneous activation of a large number of high-density tags generates a massive amount of concurrent backscattered signals to the reader, causing extremely severe channel collisions and decoding conflicts. Existing fixed-frame time-slot anti-collision algorithms are unable to adjust channel capacity in real time according to dynamic tag density, and are prone to long inventory cycles and significant decrease in recognition accuracy due to channel congestion. Summary of the Invention

[0005] To achieve the above objectives, the present invention provides the following technical solution: This invention provides an automatic inventory control method based on mobile RFID identification, including the following steps: S100: Task configuration and height control: The control terminal acquires the target inventory task and controls the height adjustment actuator to drive the antenna assembly to move vertically to the target height according to the floor height information corresponding to the target inventory task; S200: Multi-antenna time-division polling selection: The control terminal selects antenna channels with different radiation characteristics in the antenna assembly in a time-division manner through the reader and writer according to a preset polling strategy, so that each antenna channel can work independently in a time-division manner; S300: Adaptive power RF reading: The selected antenna channel transmits an RF carrier signal under the control of the reader to activate RFID tags in the target area, and the reader executes an adaptive power adjustment program during the reading process to dynamically adjust the RF output power based on tag reading feedback; S400: Edge data preprocessing: The reader collects the raw data returned by the activated tag in real time and executes the edge preprocessing program on the device. It uses preset filtering rules and signal strength thresholds to denoise and deduplicatize the raw data and select the effective tag data. S500: Real-time data comparison and synchronization: The control terminal acquires valid tag data and performs set operation comparison with the pre-stored list of tags to be stored, generates inventory results in real time including normal, missing and misplaced states, and controls the communication module to upload the inventory results to the server.

[0006] Further, in step S100, the control terminal acquires the target inventory task, including: Receive the target inventory area identifier and starting inventory level as input or obtained from the server; Determine the inventory mode corresponding to the target inventory task; wherein, the inventory mode includes a single-layer scanning mode and a layer-by-layer scanning mode; In the layer-by-layer scanning mode, when it is determined that the current layer inventory task has been completed, the control terminal automatically controls the height adjustment actuator to drive the antenna assembly to move vertically to the next target height.

[0007] Furthermore, the step of controlling the height adjustment actuator to move the antenna assembly vertically to the target height based on the floor height information corresponding to the target inventory task specifically includes the following sub-steps: S101: The control terminal calculates the target driving parameters required for the antenna assembly to reach the target height by querying a preset layer height mapping relationship based on the layer height information. S102: Output a drive command corresponding to the target drive parameters to the height adjustment actuator to initiate vertical movement; S103: Real-time acquisition of position status signals fed back by position sensors, and calculation and updating of the current height value based on the position status signals; S104: When the comparison determines that the current height value has reached the target height, or when the limit sensor receives the limit trigger signal, the height adjustment actuator is controlled to stop moving and an antenna in place status is generated.

[0008] Furthermore, the antenna assembly includes antenna channels with different radiation characteristics, including: The first antenna channel, employing high-gain wide-angle radiation characteristics, is fixedly mounted on the outer wall of the moving vehicle, with its main lobe axis oriented horizontally towards the target inventory area. This allows for large-area tag activation within the low-to-mid-level altitude range. The second antenna channel, employing narrow-beam directional radiation characteristics, is located at the active end of the height adjustment actuator, and its main radiation lobe axis shifts vertically with the height adjustment actuator. This allows for precise directional tag activation of densely stored tag areas at high altitudes after reaching the target height.

[0009] Further, in step S200, according to a preset polling strategy, the reader selects antenna channels with different radiation characteristics in the antenna assembly in a time-division multiplexing manner, specifically including: The current height value fed back by the height adjustment actuator is obtained in real time, and the storage density attribute of the current target inventory area is identified. If the current height value is within the preset low-to-mid-level height range and the storage density attribute is normal density, then the reader shuts down the second antenna channel and controls the multiple first antenna channels located on both sides of the moving vehicle to perform alternating time-division strobe. If the current height value is within a preset high-rise height range and the storage density attribute is high-density dense storage, the reader closes the first antenna channel and controls multiple second antenna channels located at the active end of the height adjustment actuator to perform time slot polling and gating, and constrains that within any gating time slot, only a single second antenna channel is allowed to be in radio frequency transmission state.

[0010] Furthermore, in step S300, the dynamic adjustment of the radio frequency output power based on tag reading feedback includes a closed-loop power optimization process executed by the control terminal: Obtain the set of tag identifiers within the current reading cycle, and perform spatial validity comparison based on the preset target location mapping relationship; When it is determined that there are interfering tags in the tag identification set that belong to adjacent physical columns or non-target inventory areas, the reader adjusts the radio frequency output power downward by a preset power attenuation step until the reading of the interfering tags is eliminated; When it is determined that there are missing tags that should have been stored but were not successfully read in the current target inventory area, the reader adjusts the radio frequency output power upward by a preset power gain step.

[0011] Furthermore, the closed-loop power optimization process also includes a degradation compensation mechanism for extreme operating conditions: Monitor the current adjustment value of the radio frequency output power and determine whether it has reached the maximum transmit power threshold allowed by the system; When the current adjustment value has reached the maximum transmit power threshold, and the missing tag is still determined after multiple consecutive reading cycles, the instruction to further increase the power is intercepted, and a scan delay compensation instruction is triggered. In response to the scanning delay compensation command, the control terminal actively reduces the horizontal movement speed of the antenna assembly within the current target area, or extends the dwell time of the current selected antenna channel in the time slot.

[0012] Further, in step S400, the execution of the edge preprocessing procedure on the device side, which uses preset filtering rules and signal strength thresholds to denoise and deduplicate the original data, specifically includes executing a data cleaning pipeline: Feature reading: Parse the raw data and extract the identification code of each response tag and the corresponding received signal strength indication feature value; Prefix filtering stage: The extracted identifier code is matched with the pre-configured target encoding prefix rules, and non-target label data that does not conform to the prefix rules is intercepted and discarded; Deduplication and aggregation stage: Within a discrete scan time period, duplicate reported data with the same identifier code are monitored and intercepted to achieve data deduplication within a single scan period; Signal denoising stage: The received signal strength indication feature value is compared with the reference strength threshold that calibrates the effective radiation depth of the target area. False tag data that are generated by signal reflection or diffraction and whose intensity is lower than the reference strength threshold are removed, and finally the effective tag data is output. For high-density storage environments, before extracting the identifier code for each response tag, the edge preprocessing procedure also performs a dynamic time-slot anti-collision control sub-step: Configure a query data frame containing the initial number of time slots and broadcast a read command; During the process of receiving the returned data, the proportion of collision time slots that fail to decode the signal and the proportion of idle time slots without tag response are continuously counted within a reading window period. When it is determined that the collision time slot ratio exceeds the preset first collision threshold, the base number of time slots for subsequent query data frames is automatically increased; When the proportion of idle time slots exceeds a preset second idle threshold, the base number of time slots for subsequent query data frames is automatically reduced until concurrency conflicts are eliminated and the extraction of the identification code is completed.

[0013] Further, in step S500, the comparison with the pre-stored list of tags to be stored, and the real-time generation of inventory results including normal, missing, and misplaced states, specifically includes performing a state determination sub-step based on set mapping: Construct the set of valid tags actually read and generated at the current target layer, as well as the set of tags to be stored retrieved from the database; Calculate the intersection of the set of tags that should be stored and the set of valid tags, and map the tags belonging to the intersection to a normal in-stock status; Calculate the first difference set between the set of tags that should be stored and the set of valid tags, map the tags belonging to the first difference set to the missing status, and generate a corresponding missing warning; Calculate the second difference set between the valid tag set and the tag set that should be stored, map the tag identifiers belonging to the second difference set to misalignment anomaly status, and generate a repositioning verification prompt; The normal in-stock status, missing status, and misaligned abnormal status, along with their corresponding tags, are rendered in a structured manner onto the interactive interface of the control terminal. The control communication module uploads the inventory results to the server, specifically implementing a multi-priority degradation and disaster recovery synchronization mechanism: The inventory results are packaged into a standard format data packet; Detect whether the first priority wireless LAN communication link is available; if available, send the data packet through the wireless LAN communication link. If the first priority wireless local area network communication link is unavailable, then the availability of the second priority wide area cellular mobile network link is checked. If available, the data packet is sent through the wide area cellular mobile network link. If all the wireless links of the above priorities are disconnected, an offline temporary storage mode is triggered, the data packets are written to a local non-volatile memory to establish a buffer queue, and an asynchronous detection thread is started in the background; when the asynchronous detection thread detects that any network link has been restored, it automatically pushes the data packets in the buffer queue to the server in batches.

[0014] On the other hand, an automatic inventory control system based on mobile RFID identification includes, The task height control module is used to acquire the target inventory task and, based on the floor height information corresponding to the target inventory task, control the height adjustment actuator to drive the antenna assembly to move vertically to the target height. The antenna time-division polling module is used to select antenna channels with different radiation characteristics in the antenna assembly in a time-division manner according to a preset polling strategy, so that each antenna channel can work independently in a time-division manner. An adaptive reading module is used to control the selected antenna channel to transmit radio frequency carrier signals under the control of the reader to activate RFID tags in the target area, and to control the reader to execute an adaptive power adjustment program during the reading process, dynamically adjusting the radio frequency output power according to the tag reading feedback; The edge data preprocessing module is used to collect the raw data returned by the activated tag in real time through the reader and writer, and execute the edge preprocessing program on the device side. It uses preset filtering rules and signal strength thresholds to denoise and deduplicatize the raw data and filter out the effective tag data. The data comparison and synchronization module is used to acquire the valid tag data through the control terminal, perform set operation comparison with the pre-stored list of tags to be stored, generate inventory results including normal, missing and misplaced states in real time, and control the communication module to upload the inventory results to the server.

[0015] Beneficial effects In use, this invention achieves efficient closed-loop control for automated inventory counting in high-density warehousing environments through task-driven vertical closed-loop height adjustment, multi-antenna time-domain isolation gating control, closed-loop adaptive power optimization, and electromechanical and hardware / software co-design of edge-side data cleaning pipelines. By utilizing closed-loop height calculation and limit redundancy protection based on absolute zero-point calibration, the spatial alignment accuracy of the antenna assembly in the vertical axis is ensured, eliminating the safety hazard of personnel falling during high-altitude manual inventory checks. Through time-division polling mutual exclusion latching of the antenna channel and closed-loop power adjustment based on RF feature stripping, the effective radiation boundary of electromagnetic waves is limited, reducing near-field coupling crosstalk and multipath reflection false reads in high-density tag groups. Relying on the dynamic time slot scheduling and edge computing pipelines such as mask filtering and hash deduplication at the reader's underlying layer, the purification and noise reduction of high-concurrency signals are completed at the data source, reducing the computing load and bus throughput pressure of the upper-level control terminal. Finally, with the state judgment based on set Boolean operations and the multi-priority degradation disaster recovery synchronization mechanism, the accurate adjudication of the inventory business logic and the complete data upload in the harsh electromagnetic environment of the warehouse are ensured. Attached Figure Description

[0016] Figure 1 This is a flowchart of the control method of the present invention; Figure 2 This is a control block diagram of the automatic inventory system of the present invention; Figure 3 This is a flowchart of the adaptive power optimization and extreme value compensation control of the present invention; Figure 4 This is a flow diagram of the edge data preprocessing pipeline of the present invention; Figure 5 This is a flowchart of the multi-priority degradation disaster recovery synchronization mechanism of the present invention. Detailed Implementation

[0017] To enable those skilled in the art to better understand the present invention, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the scope of protection of the present invention.

[0018] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this invention are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of the invention described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but includes other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0019] The present invention will now be described in further detail with reference to the accompanying drawings: Example: like Figure 1-5 As shown, an automatic inventory control method based on mobile RFID identification includes the following steps: S100: Task Configuration and Height Control: The control terminal acquires the target inventory task and, based on the floor height information corresponding to the target inventory task, controls the height adjustment actuator to drive the antenna assembly to move vertically to the target height. S200: Multi-antenna time-division polling selection: The control terminal selects antenna channels with different radiation characteristics in the antenna assembly through the reader and writer according to the preset polling strategy, so that each antenna channel can work independently in time-division. S300: Adaptive power RF reading: The selected antenna channel transmits an RF carrier signal under the control of the reader to activate RFID tags in the target area, and the reader executes an adaptive power adjustment program during the reading process to dynamically adjust the RF output power based on tag reading feedback; S400: Edge data preprocessing: The reader collects the raw data returned by the activated tag in real time and executes the edge preprocessing program on the device. It uses preset filtering rules and signal strength thresholds to denoise and deduplicatize the raw data and select the effective tag data. S500: Real-time data comparison and synchronization: The control terminal acquires valid tag data and performs set operations to compare it with the pre-stored list of tags to be stored, generating inventory results in real time that include normal, missing, and misplaced states, and controls the communication module to upload the inventory results to the server.

[0020] Furthermore, the specific implementation process of step S100 is as follows: In the specific implementation process, the specific control logic and hardware coordination mechanism for task configuration and height control in step S100 are as follows: The task configuration process begins first. The control terminal (in this embodiment, an embedded integrated industrial-grade control motherboard integrating a microprocessor, touch display, and wireless communication module) receives or acquires inventory task data through its built-in communication interface. This inventory task data can be structured parameters input by the operator via touch clicks on the graphical user interface of the touch display, or it can be instructions asynchronously retrieved from the remote document management system server via a wireless LAN module (such as a Wi-Fi module) or a wide-area cellular mobile communication module (such as a 4G / 5G module) using standard data packet formats such as JSON or XML. The received target inventory task data includes at least the target inventory area identifier (such as warehouse area code, shelf column number, and shelf number), the starting inventory level information, and the currently selected inventory mode. The inventory mode is defined in the software state machine and divided into single-layer scan mode and layer-by-layer scan mode. When the system is configured in single-layer scanning mode, the control terminal terminates the repositioning process after driving the antenna assembly to the specified single target layer height and completing RFID reading. When the system is configured in layer-by-layer scanning mode, the control terminal maintains a layer counter and inventory status flag in memory. When it is determined that the tag reading task of the current layer meets the preset termination conditions (e.g., no new EPC code is reported within 3 consecutive reading cycles, or the continuous scanning time of the current layer reaches the preset time threshold), the layer counter is automatically incremented, and the height adjustment actuator is triggered to drive the antenna assembly to automatically reposition linearly to the next target height vertically upward.

[0021] Next, the height calculation and mapping process is executed. After obtaining the floor height information corresponding to the target inventory task (e.g., the target floor is the 3rd floor), the control terminal's central processing unit calls the preset floor height mapping matrix stored in non-volatile memory (such as eMMC or Flash). This mapping matrix pre-defines the linear or non-linear mapping geometric model between the shelf floor number and the absolute spatial height, or between the floor number and the physical displacement of the actuator.

[0022] Before performing geometric model mapping and height calculation, to eliminate the physical defect of incremental position sensors losing absolute position coordinates after a system power failure and restart, the control terminal forcibly executes a mechanical absolute zero-point calibration subroutine upon initial system power-on or upon receiving the first inventory task. Specifically, the control terminal continuously outputs a reverse descent command to the motor driver, driving the telescopic lifting boom to retract downwards until the mechanical structure physically contacts and presses against the bottom lower limit sensor (normally closed microswitch). At the instant the lower limit sensor generates a hardware interrupt with a level transition, the control terminal immediately brakes the motor and forcibly clears the register value of the current hardware quadrature decoding counter to zero, thus solidifying this physical limit position as the absolute zero-point reference in the three-dimensional spatial coordinate system. All subsequent target driving parameters and actual extension displacement All operations are strictly based on this absolute zero-point reference datum for signed pulse accumulation or decrement, thereby ensuring that the absolute positioning accuracy does not accumulate drift error after multiple reciprocating lifting and lowering operations.

[0023] The central processing unit (CPU) calculates or retrieves the absolute spatial target height that the antenna assembly needs to reach by performing matrix indexing or dictionary lookup on the target layer number to be queried. Subsequently, the control terminal, based on the mechanical transmission ratio of the height adjustment actuator, the amplification factor of the linkage mechanism, and the electrical characteristics of the drive motor, determines the absolute target height value. This is converted into target drive parameters that the height adjustment actuator can directly execute. Specifically, in this embodiment, when an electrically driven push rod is used as the height adjustment actuator, the target drive parameters are represented by the total number of specific pulses required to be output by the motor driver. Or, the delay in the continuous energizing time required for a DC motor under an applied rated voltage.

[0024] If the transmission mechanism adopts a multi-stage linkage telescopic structure, the calculation formula also needs to take into account the overlap elimination constant between each stage of carbon fiber hollow rods and the mechanical slip ratio compensation coefficient, so as to accurately lock the theoretical displacement scalar required by the drive actuator.

[0025] The specific closed-loop conversion physical model formula is as follows: In the formula, The target total number of output pulses required by the motor driver; This is the absolute target height value retrieved by the system. This serves as the absolute zero reference height for the machine. The sum of the constants for eliminating the overlap of each stage of the carbon fiber hollow rod in the current extension phase of the telescopic lifting boom; This is the nominal value of the physical displacement resolution corresponding to a single pulse in the system. The mechanical slip ratio compensation coefficient is introduced to compensate for the current floor height load, and is used to offset the small step loss error in the screw drive process.

[0026] The command output and execution process then begins. The control terminal sends physical layer drive commands matching the target drive parameters to the microcontroller or motor driver (such as a motor control module integrating an H-bridge drive circuit) via its general-purpose general-purpose input / output (GPIO) pins or a dedicated serial bus interface (such as a UART bus or CAN bus). These drive commands, at the electrical level, include at least direction control signals and speed / pulse control signals. When executing an antenna assembly rise change, the control terminal pulls the direction control GPIO pin high (e.g., 3.3V or 5V), turning on the upper arm of the H-bridge in the motor driver; when executing a fall change, it pulls the pin low (e.g., 0V). Simultaneously, the control terminal's clock generator or pulse width modulation (PWM) generator begins outputting a pulse signal stream with a specific frequency and duty cycle. The motor speed is proportional to the frequency of this pulse signal, and the motor's output torque is proportional to the pulse duty cycle or the magnitude of the drive current. After receiving these weak electrical control signals, the motor driver amplifies the power through an internal power switch (such as a MOSFET), and delivers the 12V or 24V DC power supplied by the external power supply system to the drive motor winding of the height adjustment actuator in a controlled current direction and waveform amplitude. The drive motor starts to rotate and converts the rotational motion into the vertical linear extension or retraction of the telescopic lifting rod through a lead screw drive or a gear and rack drive, thereby driving the antenna assembly mounted on the mounting base at the end of the rod to move spatially along the vertical axis.

[0027] During this period, the system is in a real-time status acquisition and closed-loop feedback state. A high-precision position sensor is deployed on the height adjustment actuator; in this embodiment, a pull-wire photoelectric encoder or a Hall effect sensor array embedded at equal intervals along the longitudinal direction of the fixed section rod is specifically used. As the telescopic lifting rod moves vertically, the pull wire of the pull-wire photoelectric encoder is wound or pulled out, driving the internal photoelectric code disk to rotate, thereby generating two orthogonal pulse signals (phase A and phase B pulses) with a 90-degree phase difference. The hardware quadrature decoding counter of the control terminal performs real-time capture and edge counting of these two pulse signals at an extremely high sampling frequency (e.g., 1kHz). The direction of movement of the mechanical structure (ascending or descending) is identified by the direction of the counter's addition and subtraction. The actual extension displacement of the height adjustment actuator is calculated in real-time by multiplying the accumulated total number of pulses by the physical displacement resolution corresponding to a single pulse (e.g., 0.1mm / pulse). The central processing unit calls the geometric calculation algorithm to determine the current actual extension displacement. Adding the reference elevation of the mobile vehicle equipment compartment, the data is converted and updated to the current absolute height of the antenna assembly above the ground. This value is then dynamically written into a global variable in the kernel for the main control thread to access in real time.

[0028] Finally, the target comparison and shutdown control process begins. The interrupt service routine or high-priority comparison thread on the control terminal continuously monitors the current height value. With the target height value Real-time differential mathematical calculations are performed. To eliminate the impact of mechanical transmission inertia, system response delay, and vibration noise on control accuracy, a reasonable closed-loop convergence threshold range is set internally in the control terminal. (In this embodiment, it is set to) When the absolute value of the difference satisfies Upon arrival, the control terminal immediately determines that the antenna assembly has accurately reached the target physical layer. At this point, the control terminal immediately intercepts the current pulse output task at the software level, forcibly setting the duty cycle of the PWM signal controlling the motor driver to zero, or switching the enable pin (EN) of the motor driver to a disabled logic state. In response to this state switch, the motor driver cuts off the current to the input motor windings, or controls the H-bridge circuit to enter an energy-consumption braking state (shorting down the two ends of the motor), using the braking torque generated by the back electromotive force to overcome mechanical inertia, causing the height adjustment actuator to completely stop the vertical displacement in a very short time, achieving precise stopping without overshoot.

[0029] As a redundant safety protection mechanism at the hardware level, upper limit sensors and lower limit sensors (such as normally closed microswitches or infrared photoelectric interruptors) are rigidly installed at the upper and lower limit positions of the telescopic lifting rod's physical travel, respectively. If an algorithm anomaly or position sensor failure occurs during the comparison control process, causing the antenna assembly to exceed the normal safe operating range, when the moving part touches the upper limit sensor, the sensor's electrical circuit immediately changes from normally closed to open, directly sending a steep level transition signal to the hardware external interrupt pin (IRQ) of the control terminal. The control terminal responds to the highest priority hardware interrupt, forcibly cutting off the power supply to all drive circuits within microseconds to prevent hard collision damage to the mechanical structure. After the height adjustment actuator stops moving under normal control or safety protection, the central processing unit sets a specific bit in the system status register in memory, generating and updating the "antenna in position status" flag. This flag change serves as an event trigger signal, sending a synchronization signal to the next control stage of the software, thereby unlocking and activating the multi-antenna time-division polling selection process of the subsequent step S200.

[0030] Furthermore, the specific implementation process of step S200 is as follows: After the height adjustment actuator drives the antenna assembly to the target position and the "antenna in position" flag in the system status register is set, the control system automatically unlocks and triggers the multi-antenna time-division polling selection process in step S200. The core of this step is to dynamically allocate antenna channels with different radiation characteristics to isolate and control the spatial distribution of radio frequency energy in the time domain, thereby cutting off radio frequency electromagnetic interference in the state of multiple antennas operating concurrently.

[0031] First, the central processing unit of the control terminal executes an environment parameter identification program to read the current altitude value currently written to the kernel global variable. Simultaneously, it retrieves the shelf physical characteristic description field associated with the current target inventory task from memory, and parses the storage density attribute (such as high-density dense storage or normal density) of the current target inventory area. The control terminal will then identify the current height value. The storage density attribute is used as a boundary condition and input into a preset polling strategy state machine. The preset polling strategy state machine is pre-configured with height range thresholds (e.g., set to 2.0 meters) and time-division switching timing control parameters, and makes real-time decisions on the channel selection logic of the RF switch matrix through conditional branch statements.

[0032] In this embodiment, the antenna assembly is electrically connected to the RF output interface (such as an SMA-type external thread female connector) of the UHF RFID reader host via multiple low-loss RF coaxial cables (such as flame-retardant double-shielded coaxial cables with a characteristic impedance of 50 ohms). The reader host integrates an RF switch matrix composed of high-isolation solid-state RF switches (such as gallium arsenide field-effect transistor switches or a PIN diode switch matrix). The control terminal of this RF switch matrix is ​​directly connected to the dedicated input / output (GPIO) pins of the microprocessor inside the reader. The antenna assembly is physically deconstructed and spatially divided into a first antenna channel and a second antenna channel. The first antenna channel uses an antenna with high-gain, wide-angle radiation characteristics (such as a circularly polarized microstrip patch antenna with a half-power beamwidth of 40 degrees by 40 degrees and a spatial gain greater than 12 dBi). It is rigidly fixed on the left and right outer walls of the moving vehicle. The central axis of the main radiating lobe is horizontally distributed, with the axis direction perpendicular to the moving vehicle's direction of travel and pointing towards the middle and lower aisles of the shelf. The second antenna channel uses an antenna with narrow-beam directional radiation characteristics (such as a directional microstrip array antenna with a horizontal half-power beamwidth of 45 degrees and a vertical half-power beamwidth of 75 degrees). It is rigidly installed at the movable end of the height adjustment actuator (i.e., on the top mounting seat of the fourth section of the four-section telescopic hollow rod). The central axis of its main radiating lobe is linearly displaced in the vertical axis as the lifting rod extends and retracts, always aligned with and focused on a specific shelf level on the upper level.

[0033] When the polling strategy state machine compares and determines the current height value When the height is below the preset height threshold (2.0 meters) and the retrieved storage density attribute is normal density, the control terminal determines that the current operating condition belongs to the low-to-mid-level normal inventory area. At this time, the control terminal sends the first branch selection command to the reader host via the serial bus. The microprocessor inside the reader responds to this command by setting the control pin of the corresponding second antenna channel in the RF switch matrix to the cutoff level, so that the solid-state RF switch on this path is in a deep cutoff state, thereby physically disconnecting and isolating the RF energy transmission of the second antenna channel, making its input impedance present an extremely high impedance state (isolation greater than 30 dB). At the same time, the microprocessor inside the reader starts the internal timer, and according to the preset single-channel dwell time (e.g., set to 150 milliseconds), it outputs the selection level to the RF switch matrix in a cyclic manner through the control pin, so that the first antenna channel located on the left outer wall of the moving vehicle and the first antenna channel located on the right outer wall are alternately turned on. During conduction, the impedance of the solid-state RF switch on the corresponding path drops to an extremely low level (insertion loss less than 0.5 dB). The electromagnetic energy generated by the ultra-high frequency RF power amplifier is fully delivered to the corresponding first antenna channel through this path and radiated outward, thereby realizing tag activation and data acquisition in a large horizontal area within the mid-to-low altitude range. The alternating operation of the two first antenna channels in the time domain completely eliminates beam distortion and mutual interference caused by simultaneous transmission of the antennas on both sides.

[0034] Conversely, when the polling strategy state machine compares and determines the current height value... When the height is greater than or equal to the preset height threshold (2.0 meters) and the retrieved storage density attribute is high-density dense storage, the control terminal determines that the current working condition has switched to a high-rise dense inventory area. In this environment, the spatial distance between adjacent tags is extremely small, and any wide-angle radio frequency radiation will cause serious tag collisions and misreading of adjacent shelves. Therefore, the control terminal sends a second branch gating command to the reader host, forcibly shutting down and isolating all first antenna channels on the outer wall of the vehicle body, so that their radio frequency switches are in a completely off state. Then, the microprocessor inside the reader calls a dedicated time slot polling algorithm to introduce multiple second antenna channels installed at the active end of the height adjustment actuator into the time slot allocation queue.

[0035] The time-slot polling algorithm divides a complete scan cycle into several consecutive, non-overlapping discrete time slices (i.e., gating slots, each with a time base of 50 milliseconds). At the start of each independent gating slot, the reader's microprocessor controls the RF switch matrix to switch states through precise logic gates, executing strict mutual exclusion control logic at the hardware level. This mutual exclusion control logic enforces that during the duration of any gating slot, only one second antenna channel at the input of the RF switch matrix is ​​allowed to achieve low-loss impedance matching conduction with the output of the RF power amplifier; all other second antenna channels and the first antenna channel must be forced to remain in a deeply isolated off state. Through this hard latching mechanism of the level state machine, even if multiple narrow beam directional antennas with different fine-tuning angles are deployed at high levels, only a single narrow beam directional antenna can transmit focused radio frequency electromagnetic carriers at any microscopic instant. The radio frequency energy is precisely limited to the main lobe conical radiation space of the narrow beam antenna. This eliminates near-field electromagnetic coupling caused by the concurrency of multiple antennas and the resulting near-field intermodulation distortion from the physical and control layers, thus achieving time-division, independent, and interference-free gating of multiple antenna channels in dense high-rise environments.

[0036] Furthermore, the specific implementation process of step S300 is as follows: After the antenna channel completes precise time-division isolation and gating, the system immediately enters the adaptive power RF readout stage in step S300. The core of this step lies in using the closed-loop control logic of "software-defined RF" to dynamically optimize the RF output power based on real-time spatial electromagnetic feedback, and introducing a degradation compensation mechanism when encountering physical extreme conditions, thereby fundamentally solving the tag crosstalk and missed read problems in high-density environments.

[0037] First, the control terminal sends a start scanning command to the RFID reader host. Driven by the reader's internal RF transceiver link, the selected target antenna channel begins transmitting an RF carrier signal in the ultra-high frequency band (e.g., 860MHz to 960MHz). In the initial state, to avoid crosstalk between adjacent physical shelves caused by instantaneous high power, the reader's internal central processing unit configures the bias voltage of the RF power amplifier (PA) via a digital-to-analog converter (DAC), outputting a nominal initial RF output power (e.g., set to 25dBm, which corresponds to a nominal effective read depth of approximately 60cm in free space). After capturing the RF carrier, the passive RFID tag antenna within this effective radiation field has its internal RF front-end circuit rectify the AC energy into DC energy to activate the tag microchip. Then, using backscatter modulation technology, the stored tag identifier (EPC code) is modulated onto the reflected carrier and transmitted back to the reader.

[0038] Subsequently, the control terminal enters the core closed-loop power optimization control process. After a preset single transmit / receive cycle (typically configured to be 50 to 100 milliseconds), the reader completes baseband decoding and reports the set of tag identifiers captured during the current reading cycle to the control terminal. The control terminal's central processing unit performs a spatial legitimacy comparison operation on this tag identifier set and the pre-stored archive mapping relationship database for the current target's physical location in the system memory. This comparison operation is not merely a count, but rather uses hash tables or binary trees to precisely verify whether each extracted tag identifier legally belongs to the currently scanned target's physical coordinate domain.

[0039] Before determining power degradation adjustment, to prevent the system from misclassifying misplaced tags physically stored on the current shelf as remote electromagnetic crosstalk tags, the central processing unit inside the reader performs radio frequency feature-based stripping and classification on illegal tags identified as being outside the target area. The central processing unit extracts the Received Signal Strength Indication (RSSI) feature value and the phase angle parameter of the backscattered carrier of the illegal tag and compares it with the average RSSI reference value of legal tags in the current target layer. If the RSSI characteristic value of the illegal tag is close to or higher than the average RSSI reference value, and the phase angle deflection rate read multiple times is lower than the preset static threshold, the system determines that it is in the near-field strong coupling region and is physically defined as a 'misaligned tag in the same layer'. This type of tag does not trigger the RF power attenuation logic, but is directly transmitted to the host computer and marked as an abnormal misalignment. Conversely, if the RSSI characteristic value of the illegal tag is significantly lower than the average RSSI reference value (e.g., the difference is greater than 15dB), and the phase angle shows obvious multipath fading fluctuation characteristics, the system determines that it is a 'far-end electromagnetic crosstalk tag' caused by antenna sidelobe leakage or RF field strength overflow, and then triggers the subsequent power attenuation command.

[0040] When the comparison algorithm determines that there are additional tag codes in the current tag identification set that do not belong to the current layer or belong to an adjacent physical shelf, the system determines that the current radio frequency radiation field volume is too large, resulting in electromagnetic overflow and spatial crosstalk. At this time, the control terminal writes a power attenuation command to the power control register of the reader host. The reader adjusts the gain amplitude of the radio frequency power amplifier downward by a preset power attenuation step size (precisely configured as an discrete step value of 0.5dB or 1dB in this embodiment). After each power reduction, the system immediately initiates a new reading cycle and performs a retest. This downward recursive adjustment process continues until the identification code of the illegal crosstalk tag is completely eliminated in the receive buffer, thereby accurately converging the effective activation boundary of the radio frequency energy to the geometric boundary of the target physical layer.

[0041] Conversely, when the comparison algorithm verifies that there are missing tags in the target inventory area's tag database that should have been read but are actually missing from the tag set, the system determines that the current radio frequency field strength is insufficient to penetrate high-density media (such as dielectric loss caused by stacked paper documents) or cannot activate slightly detuned tags in the radio frequency dead zone. In this case, the control terminal sends a power gain command to the reader, which gradually increases the radio frequency output power in the same power gain step (0.5dB or 1dB) to physically expand the effective radiation volume and penetration depth of the radio frequency beam, enabling the microchip of the missing tag to obtain sufficient radio frequency threshold energy to start working.

[0042] In addition to the closed-loop power optimization process, this embodiment specifically constructs a degradation compensation mechanism for extreme operating conditions. When the control terminal outputs a power gain command, it continuously monitors the RF output power adjustment value currently fed back by the reader and compares it with the maximum transmit power threshold allowed by the system hardware (e.g., a safety upper limit set at 30dBm or the reader hardware limit of 33dBm) using hard limiting. When the actual adjustment value has reached the maximum transmit power threshold, and data verification through multiple consecutive reading cycles still determines that there are fixed missed tags, the overload protection logic at the system software level will be triggered. At this time, the control terminal actively intercepts and blocks any further control commands attempting to increase the transmit power, and instead generates and issues a scan delay compensation command.

[0043] The scanning delay compensation command manifests as a dual coordinated intervention in the electromechanical and temporal domains at the execution level. At the electromechanical level, the control terminal sends a deceleration message to the underlying driver via the motor control bus, actively reducing the horizontal or vertical linear velocity of the moving vehicle or height adjustment actuator within the current target area. This compensates for insufficient energy threshold by extending the exposure time in the physical space. At the radio frequency temporal control level, the control terminal rewrites the reader's timing control register, extending the dwell time of the currently active gating antenna channel (e.g., doubling the single-channel polling dwell time from the conventional 50 milliseconds to over 150 milliseconds). This cross-dimensional compensation mechanism allows the energy storage capacitors inside the RFID tag microchip, located at the electromagnetic edge or severely affected by multipath fading, to utilize a longer time window for time integration and charge accumulation of weak radio frequency energy. This significantly increases the probability of radio frequency activation and successful backscattering of a single stubborn tag under extreme conditions, from a mathematical expectation perspective, without exceeding the radio frequency hardware emission limits, ensuring the absolute integrity of the read data at the underlying level.

[0044] Furthermore, the specific implementation process of step S400 is as follows: After completing adaptive power RF reading and acquiring the weak backscattered signal from the tag, the system enters the edge data preprocessing stage in step S400. The core of this step lies in fully utilizing the underlying hardware computing power of the RFID reader host (such as the built-in industrial-grade ARM processor and Linux operating system) to directly execute high-concurrency signal decomposition and cleaning pipelines at the device end (edge ​​side) close to the data source. This not only eliminates concurrency conflicts in high-density storage environments at the physical layer but also, without consuming the CPU resources of the upper-level control terminal (i.e., the Android all-in-one machine), relies on pure mathematical rules and logical operations to purify massive amounts of redundant raw RF signals into high-value, effective tag data.

[0045] First, to address the concurrent response issue arising from the simultaneous activation of a massive number of tags in a high-density storage environment, the RF baseband processing chip inside the reader must execute a discrete-time dynamic time-slot anti-collision control sub-step before extracting the tag identification code. The reader broadcasts a query data frame (Query command) into the RF radiation field, which physically carries an initial frame length parameter. (e.g., initial allocation) At this point, the current frame is hard-divided into... Each RFID tag has 16 discrete time slices (i.e., 16 gated time slots). The random number generator inside each active RFID tag generates a number ranging from 0 to... The data is a random integer between the specified ranges and is sent back to the reader via backscattering within its corresponding gating time slot.

[0046] During the physical window period for receiving the returned data frame, the reader's baseband decoder performs Fast Fourier Transform and Cyclic Redundancy Check (CRC) on the RF echo signal of each gating time slot. If the CRC check is successful in a certain time slot, it is recorded as a successful time slot. ; If the baseband signal captures an RF echo within the same gating time slot, but the baseband demodulator detects an illegal phase flip (i.e., an edge transition that does not conform to the encoding rules) at the symbol edge during sampling and decision-making for FM0 baseband encoding or Miller subcarrier encoding, or if the preamble correlation peak detection fails, accompanied by a final cyclic redundancy check (CRC-16) decoding failure, the reader's underlying hardware state machine will trigger the collision flag register to be set. The system strictly relies on the waveform distortion characteristics of this physical layer to accurately determine this time slot as a collision time slot caused by the electromagnetic superposition of multiple tag signals. This eliminates data corruption caused purely by spatial Gaussian white noise.

[0047] If the baseband detector fails to capture an RF envelope higher than the noise floor, it is recorded as an idle time slot. After a complete query data frame ends, the reader's underlying microprocessor immediately initiates state machine calculations: calculating the collision time slot ratio. and idle time slot ratio When the arithmetic logic unit determines the collision time slot ratio When the first collision threshold is exceeded, it indicates that the tag density in the current radio frequency field is much greater than the available time slot resources. The reader will automatically increase the base number of time slots using an algorithm during the next frame broadcast (e.g., by executing...). Or increase in a stepwise manner, with a maximum limit of Conversely, when determining the proportion of idle time slots... Exceeding the preset second idle threshold (e.g.) When this happens, the reader automatically executes... The frame length of subsequent queries is reduced. Through this closed-loop negative feedback optimization calculation based on statistical probability, the system dynamically approaches the optimal channel capacity matching point within milliseconds, thereby completely eliminating concurrent conflicts at the physical layer.

[0048] After the anti-collision mechanism resolves channel congestion, the reader enters the feature extraction and data cleaning pipeline stage. For each successfully decoded single-sideband signal, the baseband processor extracts the complete 96-bit or 128-bit Electronic Product Code (EPC). Simultaneously, the analog-to-digital converter (ADC) captures the baseband RF envelope level at the moment the demodulator receives the data frame, quantizes it into a Received Signal Strength Indication Feature (RSSI) in dBm, and adds a high-precision hardware timestamp to the data set. These three sets of variables together constitute a raw data tuple.

[0049] The reader initiates a prefix filtering phase in memory. The reader's central processing unit performs a bitwise masking AND operation on the extracted identification code and the target encoding prefix rule (such as a 24-bit manufacturer identification code plus a category feature code representing a specific company asset or archive) pre-configured in non-volatile memory. If the operation result does not match, the system determines that the tag is a non-system tag that has mistakenly entered the inventory area (such as an access card with an RFID chip carried by a staff member or a logistics packaging label). The memory management unit then intercepts the original data tuple and directly releases the memory pointer it occupies, cutting off the inflow of invalid data from the very beginning.

[0050] The surviving legitimate data stream enters the deduplication and aggregation stage. In actual high-speed inventory operations, because the antenna transmits in continuous wave mode, a single target tag located in the RF core area is often repeatedly activated and successfully read dozens of times within a discrete scan time period (e.g., 500 milliseconds), generating a massive number of redundant tuples. The reader dynamically constructs a frequently accessed hash table in RAM. For each newly incoming legitimate data tuple, the processor uses its identifier code (EPC) as the hash key to perform hash addressing. If the key does not exist in the hash table, it is inserted as a new node; if it is determined that the key already exists (i.e., it belongs to the duplicate reported data within the same period), the system executes data aggregation logic, comparing only the RSSI feature value of the current tuple with the RSSI peak value already stored in the hash table. If the current value is larger, the peak value is overwritten and updated, and the timestamp of the most recent read is refreshed. Then, the duplicate redundant tuple is discarded from the stack. This deduplication and aggregation stage hard-compresses the originally exponentially growing concurrent raw read stream into a one-to-one tag feature dictionary in the underlying hardware RAM.

[0051] The signal denoising stage addresses the issue of radio frequency (RF) physical characteristics. Because warehouse shelving typically contains numerous metal baffles, strong multipath effects can cause RF energy from the antenna's back or side lobes to be reflected and diffracted, inadvertently activating detached tags physically outside the target area (such as adjacent rear shelves). To eliminate such spurious reads, the reader invokes a preset baseline strength threshold. .

[0052] This benchmark strength threshold It is not a fixed constant, but rather the extreme value of the distance attenuation boundary calculated in real time by the central processing unit based on the logarithmic distance path loss model. Its specific mathematical expression is: In the formula, To be at the reference distance The baseline value of the received signal strength measured at a distance of 1 meter (usually taken as the reference value); This is the preset effective radiation boundary distance of the current target layer's physical geometry; This is the path loss index for a specific high-density storage environment (typical values ​​are between 2.5 and 3.5). This is the system hardware attenuation constant after taking into account coaxial cable attenuation and RF switch matrix insertion loss.

[0053] The processor iterates through each independent tag node in the hash table after deduplication and aggregation, and updates the peak RSSI eigenvalue with the baseline intensity threshold. A rigorous floating-point number comparison is performed. Any node below the strength threshold is determined by the physical electromagnetic wave attenuation law to be outside the effective depth core region of the antenna's near-field radiation. This node is then identified as a reflection false read tag and removed from the hash table.

[0054] Furthermore, the specific implementation process of step S500 is as follows: After completing edge data preprocessing and obtaining high-confidence valid tag data, the system enters the final closed-loop stage, namely the real-time data comparison and synchronization process in step S500. The core of this step lies in using a rigorous discrete mathematical set mapping mechanism to complete the logical decision-making of business status locally on the device, and relying on a multi-priority degradation disaster recovery synchronization mechanism to ensure the absolute security and complete upload of inventory data in complex warehouse network environments.

[0055] First, the central processing unit of the control terminal executes a state determination sub-step based on set mapping. The control terminal receives valid tag data from the RFID reader via a serial bus or internal network socket, and constructs the set of valid tags actually read at the current target layer in the dynamic random access memory (DRAM) based on the tag's identification code (EPC encoding). Meanwhile, the control terminal retrieves the list of files that should be stored at the current target level in the system ledger by calling a local embedded database (such as an SQLite database) or a data dictionary previously cached from the server, and structures it into a set of tags to be stored. .

[0056] After constructing the two data sets mentioned above, the arithmetic logic unit inside the central processing unit performs multi-dimensional set difference and intersection Boolean operations.

[0057] Specifically, the central processing unit first calculates the set of tags that should be stored. With valid tag set The intersection (i.e.) The tags belonging to this intersection, in terms of physical logic, mean "that they should exist on the books and are actually read from the database." The system then maps the business attribute fields of these tags to a normal in-database state. Next, the central processing unit calculates the set of tags that should exist. Relative to the effective tag set The first difference set (i.e. Tags belonging to the first difference set indicate that they "should exist on paper but were not actually captured by the RF antenna." The system maps these to a missing state and generates a corresponding missing alert object in memory. Finally, the central processing unit calculates the set of valid tags. Relative to the set of tags to be stored The second difference set (i.e. The tag belonging to the second difference set means "actually read but not belonging to the target layer in the accounting records". The system determines that there has been a physical storage location offset, maps it to a misalignment abnormal state, and generates a repositioning verification prompt.

[0058] After completing the aforementioned set operations, the graphics processing unit (GPU) of the control terminal converts the normal in-stock state, missing state, and misaligned abnormal state, along with their corresponding labels, into a structured user interface data dictionary. This data is then rendered onto the control terminal's interactive interface (such as a 15.6-inch touchscreen display) via a low-voltage differential signaling (LVDS) interface. Data entries in different mapping states are assigned different visual status bits (e.g., a green background for normal states, a red background for missing states, and a yellow background for misaligned states) to visualize the results of the underlying algorithm's computation.

[0059] After the comparison results are generated, the system seamlessly integrates a multi-priority degradation and disaster recovery synchronization mechanism. The microprocessor of the control terminal first serializes the inventory results, timestamps, target shelf coordinates, and operator authentication information obtained from the above calculations, and encapsulates them into standard format data packets conforming to the RESTful architecture (such as JSON format strings). Subsequently, the network layer protocol stack initiates link availability detection. The system first checks the association status and IP address allocation of the highest priority wireless LAN communication link (such as IEEE 802.11b / g / n Wi-Fi modules). If the link is detected to be available and can successfully resolve the server domain name, a TCP connection is established and data packets are sent via the wireless LAN communication link using an HTTP POST request.

[0060] If the network socket returns a connection timeout, the first-priority Wi-Fi communication link is determined to be unavailable. The control terminal immediately wakes up the second-priority wide-area cellular network link (such as an external 4G / 5G LTE communication module) via the internal bus. After confirming that the cellular network module has successfully registered with the operator's base station, the protocol stack switches the physical network card route and sends data packets through the wide-area cellular network link, thereby overcoming the limitations of Wi-Fi signal coverage blind spots in some large reinforced concrete warehouses. For the most severe extreme network outage scenario—when all the aforementioned priority wireless links experience physical disconnection or network layer blockage—the control terminal's operating system kernel will intercept all network layer write operations and trigger an offline temporary storage mode. In this mode, the control terminal persistently saves the serialized JSON data packets to an independent cache queue established in local non-volatile memory (such as eMMC flash memory chips) or a local database via append operations, ensuring data is not lost even when the device loses power. Simultaneously, the control terminal runs in the background and starts an asynchronous detection thread. This asynchronous detection thread sends ICMP echo requests (Ping packets) to the gateway or external network core server at preset time intervals (e.g., every 30 seconds). When this asynchronous detection thread detects that any network link has been restored and the network quality meets the transmission threshold, the system triggers a callback function, automatically retrieving all offline data packets accumulated in the cache queue from the non-volatile memory and synchronizing them to the cloud server in batches. Only after the server returns an HTTP 200 OK confirmation code is the local cache queue released and cleared.

[0061] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions will not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. An automatic inventory control method based on mobile RFID identification, characterized in that, Includes the following steps: S100: Task Configuration and Height Control: The control terminal acquires the target inventory task and, based on the floor height information corresponding to the target inventory task, controls the height adjustment actuator to drive the antenna assembly to move vertically to the target height. S200: Multi-antenna time-division polling selection: The control terminal selects antenna channels with different radiation characteristics in the antenna assembly in a time-division manner through the reader and writer according to a preset polling strategy, so that each antenna channel can work independently in a time-division manner; S300: Adaptive power RF reading: The selected antenna channel transmits an RF carrier signal under the control of the reader to activate RFID tags in the target area, and the reader executes an adaptive power adjustment program during the reading process to dynamically adjust the RF output power based on tag reading feedback; S400: Edge data preprocessing: The reader collects the raw data returned by the activated tag in real time and executes the edge preprocessing program on the device. It uses preset filtering rules and signal strength thresholds to denoise and deduplicatize the raw data and select the effective tag data. S500: Real-time data comparison and synchronization: The control terminal acquires valid tag data and performs set operation comparison with the pre-stored list of tags to be stored, generates inventory results in real time including normal, missing and misplaced states, and controls the communication module to upload the inventory results to the server.

2. The automatic inventory control method based on mobile RFID identification according to claim 1, characterized in that, In step S100, the control terminal acquires the target inventory task, including: Receive the target inventory area identifier and starting inventory level as input or obtained from the server; Determine the inventory mode corresponding to the target inventory task; wherein, the inventory mode includes a single-layer scanning mode and a layer-by-layer scanning mode; In the layer-by-layer scanning mode, when it is determined that the current layer inventory task has been completed, the control terminal automatically controls the height adjustment actuator to drive the antenna assembly to move vertically to the next target height.

3. The automatic inventory control method based on mobile RFID identification according to claim 2, characterized in that, The step of controlling the height adjustment actuator to move the antenna assembly vertically to the target height based on the floor height information corresponding to the target inventory task specifically includes the following sub-steps: S101: The control terminal calculates the target driving parameters required for the antenna assembly to reach the target height by querying a preset layer height mapping relationship based on the layer height information. S102: Output a drive command corresponding to the target drive parameters to the height adjustment actuator to initiate vertical movement; S103: Real-time acquisition of position status signals fed back by position sensors, and calculation and updating of the current height value based on the position status signals; S104: When the comparison determines that the current height value has reached the target height, or when the limit sensor receives the limit trigger signal, the height adjustment actuator is controlled to stop moving and an antenna in place status is generated.

4. The automatic inventory control method based on mobile RFID identification according to claim 3, characterized in that, The antenna assembly includes antenna channels with different radiation characteristics, including: The first antenna channel, employing high-gain wide-angle radiation characteristics, is fixedly installed on the outer wall of the mobile vehicle body, with its main radiation lobe axis oriented horizontally toward the target inventory area, for performing tag activation over a wide area in the mid-to-low layer height range. The second antenna channel, employing narrow-beam directional radiation characteristics, is located at the active end of the height adjustment actuator, and its main radiation lobe axis shifts vertically with the height adjustment actuator. This allows for precise directional tag activation of densely stored tag areas at high altitudes after reaching the target height.

5. The automatic inventory control method based on mobile RFID identification according to claim 4, characterized in that, In step S200, according to a preset polling strategy, the reader selects antenna channels with different radiation characteristics in the antenna assembly in a time-division multiplexing manner, specifically including: The current height value fed back by the height adjustment actuator is obtained in real time, and the storage density attribute of the current target inventory area is identified. If the current height value is within the preset low-to-mid-level height range and the storage density attribute is normal density, then the reader shuts down the second antenna channel and controls the multiple first antenna channels located on both sides of the moving vehicle to perform alternating time-division strobe. If the current height value is within a preset high-rise height range and the storage density attribute is high-density dense storage, the reader closes the first antenna channel and controls multiple second antenna channels located at the active end of the height adjustment actuator to perform time slot polling and gating, and constrains that within any gating time slot, only a single second antenna channel is allowed to be in radio frequency transmission state.

6. The automatic inventory control method based on mobile RFID identification according to claim 5, characterized in that, In step S300, the dynamic adjustment of the radio frequency output power based on tag reading feedback includes a closed-loop power optimization process executed by the control terminal: Obtain the set of tag identifiers within the current reading cycle, and perform spatial validity comparison based on the preset target location mapping relationship; When it is determined that there are interfering tags in the tag identification set that belong to adjacent physical columns or non-target inventory areas, the reader adjusts the radio frequency output power downward by a preset power attenuation step until the reading of the interfering tags is eliminated; When it is determined that there are missing tags that should have been stored but were not successfully read in the current target inventory area, the reader adjusts the radio frequency output power upward by a preset power gain step.

7. The automatic inventory control method based on mobile RFID identification according to claim 6, characterized in that, The closed-loop power optimization process also includes a degradation compensation mechanism for extreme operating conditions: Monitor the current adjustment value of the radio frequency output power and determine whether it has reached the maximum transmit power threshold allowed by the system; When the current adjustment value has reached the maximum transmit power threshold, and the missing tag is still determined after multiple consecutive reading cycles, the instruction to further increase the power is intercepted, and a scan delay compensation instruction is triggered. In response to the scanning delay compensation command, the control terminal actively reduces the horizontal movement speed of the antenna assembly within the current target area, or extends the dwell time of the current selected antenna channel in the time slot.

8. The automatic inventory control method based on mobile RFID identification according to claim 7, characterized in that, In step S400, the execution of the edge preprocessing procedure on the device side, which uses preset filtering rules and signal strength thresholds to denoise and deduplicate the original data, specifically includes executing a data cleaning pipeline: Feature reading: Parse the raw data and extract the identification code of each response tag and the corresponding received signal strength indication feature value; Prefix filtering stage: The extracted identifier code is matched with the pre-configured target encoding prefix rules, and non-target label data that does not conform to the prefix rules is intercepted and discarded; Deduplication and aggregation stage: Within a discrete scan time period, duplicate reported data with the same identifier code are monitored and intercepted to achieve data deduplication within a single scan period; Signal denoising stage: The received signal strength indication feature value is compared with the reference strength threshold that calibrates the effective radiation depth of the target area. False tag data that are generated by signal reflection or diffraction and whose intensity is lower than the reference strength threshold are removed, and finally the effective tag data is output. For high-density storage environments, before extracting the identifier code for each response tag, the edge preprocessing procedure also performs a dynamic time-slot anti-collision control sub-step: Configure a query data frame containing the initial number of time slots and broadcast a read command; During the process of receiving the returned data, the proportion of collision time slots that fail to decode the signal and the proportion of idle time slots without tag response are continuously counted within a reading window period. When it is determined that the collision time slot ratio exceeds the preset first collision threshold, the base number of time slots for subsequent query data frames is automatically increased; When the proportion of idle time slots exceeds a preset second idle threshold, the base number of time slots for subsequent query data frames is automatically reduced until concurrency conflicts are eliminated and the extraction of the identification code is completed.

9. The automatic inventory control method based on mobile RFID identification according to claim 8, characterized in that, In step S500, the comparison with the pre-stored list of tags to be stored, and the generation of inventory results containing normal, missing, and misplaced states in real time, specifically includes performing a state determination sub-step based on set mapping: Construct the set of valid tags actually read and generated at the current target layer, as well as the set of tags to be stored retrieved from the database; Calculate the intersection of the set of tags that should be stored and the set of valid tags, and map the tags belonging to the intersection to a normal in-stock status; Calculate the first difference set between the set of tags that should be stored and the set of valid tags, map the tags belonging to the first difference set to the missing status, and generate a corresponding missing warning; Calculate the second difference set between the valid tag set and the tag set that should be stored, map the tag identifiers belonging to the second difference set to misalignment anomaly status, and generate a repositioning verification prompt; The normal in-stock status, missing status, and misaligned abnormal status, along with their corresponding tags, are rendered in a structured manner to the interactive interface of the control terminal. The control communication module uploads the inventory results to the server, specifically executing a multi-priority degradation disaster recovery synchronization mechanism: The inventory results are packaged into a standard format data packet; Detect whether the first priority wireless LAN communication link is available; if available, send the data packet through the wireless LAN communication link. If the first priority wireless local area network communication link is unavailable, then the availability of the second priority wide area cellular mobile network link is checked. If available, the data packet is sent through the wide area cellular mobile network link. If all the wireless links of the above priorities are disconnected, an offline temporary storage mode is triggered, the data packets are written to a local non-volatile memory to establish a buffer queue, and an asynchronous detection thread is started in the background; when the asynchronous detection thread detects that any network link has been restored, it automatically pushes the data packets in the buffer queue to the server in batches.

10. An automatic inventory control system based on mobile RFID identification, comprising an automatic inventory control method based on mobile RFID identification according to any one of claims 1-9, characterized in that, The system includes, The task height control module is used to acquire the target inventory task and, based on the floor height information corresponding to the target inventory task, control the height adjustment actuator to drive the antenna assembly to move vertically to the target height. The antenna time-division polling module is used to select antenna channels with different radiation characteristics in the antenna assembly in a time-division manner according to a preset polling strategy, so that each antenna channel can work independently in a time-division manner. An adaptive reading module is used to control the selected antenna channel to transmit radio frequency carrier signals under the control of the reader to activate RFID tags in the target area, and to control the reader to execute an adaptive power adjustment program during the reading process, dynamically adjusting the radio frequency output power according to the tag reading feedback; The edge data preprocessing module is used to collect the raw data returned by the activated tag in real time through the reader and writer, and execute the edge preprocessing program on the device side. It uses preset filtering rules and signal strength thresholds to denoise and deduplicatize the raw data and filter out the effective tag data. The data comparison and synchronization module is used to acquire the valid tag data through the control terminal, perform set operation comparison with the pre-stored list of tags to be stored, generate inventory results including normal, missing and misplaced states in real time, and control the communication module to upload the inventory results to the server.

Images