RFID high-gain phased-array antenna array

Abstract

The invention relates to the technical field of wireless communication, in particular to an RFID high-gain phased-array antenna array. The system comprises an RFID tag module, a direction switching module and a beam optimization module. A coordinate system is constructed through the RFID tag module, the three-dimensional position of a target tag is solved in real time, the three-dimensional position of the target tag is obtained through the direction switching module, a phased-array antenna is formed through the power divider and the 3dB bridge, the power divider is used for carrying out equal-amplitude distribution on input radio-frequency signals, the phase difference characteristic of the 3dB bridge is combined, the wave beam pointing direction of a phased array is changed, and the phase difference characteristic of the 3dB bridge is combined, so that the three-dimensional position of the target tag is obtained. And the beam optimization module establishes a neural network model, dynamically adjusts the pointing beam of the phased array, evaluates the current beam performance by taking the tag identification success rate as an index, and performs feedback optimization on the beam, so that the signal coverage area is adjusted according to the retail scene requirement, and the tag identification rate is improved.

Description

Technical Field

[0001] The present invention relates to the field of wireless communication technology, and in particular to an RFID high-gain phased array antenna array. Background Art

[0002] In the retail industry, RFID technology has been widely used in merchandise management, inventory monitoring, and other aspects. Traditional RFID antennas, mostly single antennas or simple arrays, suffer from low gain, fixed directional patterns, and an inability to adapt flexibly to complex retail environments. For example, in retail environments with densely packed shelves and numerous metal obstacles, traditional antennas are prone to insufficient signal coverage, low tag read rates, and severe multipath interference.

[0003] To address these issues, phased array antenna technology has been gradually introduced. By controlling the phase and amplitude of each antenna element, phased array antennas achieve flexible beam scanning and shaping, effectively improving signal coverage and tag recognition rates. However, existing phased array antennas are large, complex, and expensive, making them difficult to meet the requirements of retail scenarios for antenna miniaturization and low cost. To combine RFID and phased array antenna technology, we optimize the antenna structure and control strategy to achieve antenna miniaturization, high gain, circularly polarized radiation, and flexible beam direction switching. Furthermore, we dynamically adjust beam signal parameters based on the environment near the target tag to reduce multipath interference. Therefore, we propose an RFID high-gain phased array antenna array. Summary of the Invention

[0004] The purpose of this invention is to solve the problems that traditional RFID antennas are mostly single antennas or simple array antennas, which have low gain, fixed radiation patterns, and cannot flexibly adapt to complex retail environments. In order to combine RFID technology with phased array antenna technology, by optimizing the antenna structure and control strategy, the antenna can be miniaturized, with high gain, circularly polarized radiation, and flexible switching of beam direction. In addition, the beam signal parameters can be dynamically adjusted according to the environment near the target tag to reduce multipath interference.

[0005] To achieve the above objectives, the present invention provides an RFID high-gain phased array antenna array, comprising an RFID tag module, a direction switching module, and a beam optimization module;

[0006] The RFID tag module constructs a three-dimensional reference coordinate system through the reference tag, and uses the time difference between the radio frequency signal reaching the reference tag and the target tag to calculate the three-dimensional position of the target tag in real time;

[0007] The direction switching module obtains the three-dimensional position of the target tag, forms a phased array antenna through a power splitter and a 3dB bridge, uses the power splitter to distribute the input RF signal in equal amplitude, and combines the phase difference characteristics of the 3dB bridge to change the beam direction of the phased array to point to the position of the target tag;

[0008] The beam optimization module collects environmental data and radio frequency signal data of the shelf where the target tag is located, uses the environmental data and radio frequency characteristic data as input, establishes a neural network model, outputs the optimal beam parameters pointing to the target tag, dynamically adjusts the pointing beam of the phased array, evaluates the current beam performance based on the tag recognition success rate, and performs feedback optimization on the beam.

[0009] Compared with the prior art, the present invention has the following beneficial effects:

[0010] 1. This high-gain RFID phased array antenna array uses RFID tag modules to establish a coordinate system. It uses the time difference between the RF signal reaching the reference tag and the target tag to calculate the target tag's three-dimensional position in real time. The direction switching module obtains the target tag's three-dimensional position. A power divider and a 3dB bridge are used to form a phased array antenna. The power divider distributes the input RF signal with equal amplitude. Combined with the phase difference characteristics of the 3dB bridge, the signal phase difference is adjusted using the 3dB bridge, enabling circular polarization radiation and flexible beam direction switching. Circular polarization enhances the antenna's ability to receive tags from different directions. The beam direction switching function adjusts the signal coverage area according to retail scenario requirements, improving tag recognition rates.

[0011] 2. The beam optimization module collects environmental data and RF signal data of the shelf where the target tag is located. Using this data and RF characteristic data as input, it establishes a neural network model, outputs the optimal beam parameters pointing to the target tag, and dynamically adjusts the pointing beam of the phased array. Using the tag recognition success rate as an indicator, it evaluates the current beam performance and performs feedback optimization on the beam. This allows the module to dynamically adjust the beam signal parameters based on the environment near the target tag, reduce multipath interference, and improve anti-interference capabilities and system flexibility.

[0012] As a further improvement of this technical solution, the RFID tag module fixes reference tags at four fixed points on the shelf where the target tag is located to construct a three-dimensional coordinate reference system. Each reference tag is equipped with a high-precision clock to support synchronous trigger signal transmission.

[0013] As a further improvement of the present technical solution, the RFID tag module synchronizes the signal transmission time of each reference tag, measures the time difference between the radio frequency signal and the target tag, uses the time difference to construct a nonlinear equation group, and adopts the least squares method to iteratively solve the three-dimensional coordinates of the target tag.

[0014] The beneficial effect of adopting the above further improvement is that the four reference tags (such as points A, B, C, and D installed at the top corners of the shelf) form a fixed three-dimensional coordinate system (such as with A as the origin, AB as the X-axis, AD as the Y-axis, and the vertical direction as the Z-axis), providing an absolute coordinate reference for the target tag (such as the cargo tag on the shelf), avoiding the cumulative error of traditional relative positioning (such as node-by-node ranging);

[0015] If the shelf is physically displaced (such as a warehouse robot moving the shelf), the coordinate system can be calibrated in real time by verifying the relative distance between reference tags (such as the known fixed distances of AB, AC, and AD) to ensure the stability of the positioning reference.

[0016] As a further improvement of the present technical solution, the direction switching module distributes the input RF signal to the phased array with equal amplitude through a power divider, wherein the connection between the power divider and the bridge is as follows:

[0017] The power splitter receives the radio frequency signal of the RFID reader, the door-entry output of the power splitter is connected to the input end of the bridge, and the exit output of the power splitter is connected to the isolation end of the bridge.

[0018] The beneficial effect of adopting the above further improvement is that the power splitter distributes the power of the reader's RF signal to the two paths of the phased array (entry direction and exit direction), ensuring that the amplitude of the two signals is the same. This lays the foundation for the subsequent phase processing of the bridge, avoids signal distortion or phase error caused by power differences, and improves the overall stability of the system.

[0019] Precise phase control can optimize the beamforming effect of the phased array, reduce the sidelobe signal strength, reduce the impact of multipath interference in the environment (such as wall reflection and metal object scattering) on ​​the tag signal, and improve the signal-to-noise ratio in the mainlobe direction, thereby improving the reading reliability and positioning accuracy of the RFID system.

[0020] As a further improvement of the present technical solution, the direction switching module changes the excitation amplitude of the corresponding phased array by controlling the switching states of different output ports of the power divider, adjusts the phase difference between the two signals through a 3dB bridge, and switches the beam direction of the phased array.

[0021] The beneficial effect of adopting the above further improvement is that the switching state of the power divider can control the signal power of each antenna unit (for example, closing some ports can reduce the excitation amplitude of the corresponding array). Combined with the phase difference of the bridge, dynamic fine-tuning of the beam shape (for example, compressing the main lobe width and suppressing the side lobes) can be achieved.

[0022] By switching the beam direction, the signal can be focused on tags in a specific area, suppressing interfering tags in other areas (such as non-target tags within the coverage range of the same reader), improving signal resolution in a multi-tag environment, and reducing the probability of collision.

[0023] As a further improvement of the present technical solution, the direction switching module is provided with an electromagnetic bandgap structure and a high-impedance surface between adjacent arrays in the phased array antenna to suppress the mutual coupling between the arrays. At the same time, the isolation port of the 3dB bridge is connected to a 50-ohm matching resistor to absorb the reflected power.

[0024] The beneficial effect of adopting the above further improvement is that by introducing the bandgap effect, electromagnetic waves of specific frequencies are prevented from propagating between adjacent array elements. This is equivalent to building an "electromagnetic barrier" between the elements, suppressing near-field coupling. By utilizing the high impedance characteristics, the currents of adjacent elements are phase-opposite, canceling the electromagnetic field generated by mutual coupling. In particular, the mutual coupling coefficient can be reduced by more than 10dB near the resonant frequency.

[0025] Mutual coupling can cause the excitation amplitude and phase of array elements to deviate from the designed values, resulting in beam pointing deviation, main lobe widening, and side lobe elevation. By suppressing mutual coupling, it is ensured that each element radiates signals according to the ideal phase difference, maintaining beamforming accuracy.

[0026] Ideally, the isolated port of a 3dB bridge (such as the fourth port of a hybrid network) should have no signal output. However, in practice, if the load impedance does not match the characteristic impedance (50Ω), a reflected signal will return to the input port, causing a deterioration in the standing wave ratio and signal distortion.

[0027] Connecting a 50Ω resistor puts the isolated port into a matched state, converting reflected power into heat absorption and ensuring the integrity of the input signal. For example, when the bridge isolation is 20dB, the matching resistor can absorb more than 99% of the leakage power, preventing it from interfering with the main signal path.

[0028] As a further improvement of this technical solution, the beam optimization module generates radio frequency propagation data under different environments through electromagnetic simulation software, records the corresponding optimal beam parameters, constructs a training set, and trains the neural network model.

[0029] As a further improvement of this technical solution, the beam optimization module uses a convolutional layer to extract environmental image feature data, extracts radio frequency feature data through a fully connected layer, splices the environmental image feature data and the radio frequency feature data, and generates beam parameters through multiple layers of fully connected layers.

[0030] The beneficial effect of adopting the above further improvements is that electromagnetic simulation software (such as HFSS and CST) can simulate the RF propagation characteristics under different scenarios (such as multiple obstacles indoors, metal reflective surfaces, and dynamic human movement), and generate data sets containing parameters such as path loss, multipath effect, and phase delay;

[0031] The environmental image features (spatial dimension) and the RF features (electromagnetic dimension) are fused through the splicing layer, breaking the limitation of the separation of environmental modeling and signal processing in traditional methods. The multi-layer fully connected layer fits the complex nonlinear relationship between beam parameters (such as phase difference, amplitude weight) and environmental-RF characteristics through activation functions (ReLU, Sigmoid), which is superior to traditional beamforming algorithms based on analytical formulas.

[0032] As a further improvement of the present technical solution, the beam optimization module records the number of responses of each tag through the RFID reader, counts the proportion of tags that respond successfully, and performs multiple repeated tests to take the average value to eliminate random errors.

[0033] As a further improvement of the present technical solution, the beam optimization module divides the tag recognition success rate into four evaluation levels: excellent, good, average, and poor, and formulates corresponding optimization suggestions.

[0034] The beneficial effect of adopting the above further improvement is that by counting the proportion of successfully responding tags (such as the proportion of tags correctly identified in the total number of tests), the RFID system performance is converted into a quantifiable numerical indicator (such as a success rate of 95% or 80%), thus avoiding misjudgments caused by manual observation;

[0035] For "good" grade systems, there is no need to invest in high-cost hardware upgrades. Performance can be improved and operation and maintenance costs can be reduced by simply adjusting software parameters (such as optimizing anti-collision algorithms and extending read and write times). In large-scale deployment scenarios (such as smart warehousing), step-by-step optimization can be performed according to regional levels, prioritizing "poor" areas and then processing "general" areas to improve resource utilization efficiency.

[0036] In addition to the above-described objects, features and advantages, the present invention has other objects, features and advantages. The present invention will be further described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 It is a schematic diagram of the overall process of the present invention;

[0038] Figure 2 Schematic diagram of the connection between the power divider and the bridge of the present invention;

[0039] Figure 3 Schematic diagram of phased array beam switching of the present invention;

[0040] Figure 4 Schematic diagram of circular polarization switching of the present invention;

[0041] Figure 5 This is a schematic diagram of the feature splicing and fusion of the present invention.

[0042] The meaning of each number in the figure is:

[0043] 100, RFID tag module; 200, direction switching module; 300, beam optimization module. DETAILED DESCRIPTION

[0044] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making any creative efforts shall fall within the scope of protection of the present invention.

[0045] Currently, traditional RFID antennas are mostly single antennas or simple array antennas, which have low gain, fixed radiation patterns, and cannot flexibly adapt to complex retail environments. In order to combine RFID technology with phased array antenna technology, by optimizing the antenna structure and control strategy, the antenna can be miniaturized, with high gain, circularly polarized radiation, and flexible switching of beam directions. In addition, the beam signal parameters can be dynamically adjusted according to the environment near the target tag to reduce multipath interference.

[0046] Therefore, the present invention proposes to construct a coordinate system through the RFID tag module, calculate the three-dimensional position of the target tag in real time, obtain the three-dimensional position of the target tag through the direction switching module, form a phased array antenna through a power divider and a 3dB bridge, use the power divider to distribute the input RF signal with equal amplitude, combine the phase difference characteristics of the 3dB bridge, change the beam direction of the phased array to point to the position of the target tag, and collect the environmental data and RF signal data of the shelf where the target tag is located. Establish a neural network model, dynamically adjust the directional beam of the phased array, evaluate the current beam performance with the tag recognition success rate as an indicator, and perform feedback optimization on the beam.

[0047] The details are as follows:

[0048] See also Figure 1 As shown, the present invention provides an RFID high-gain phased array antenna array, including an RFID tag module 100, a direction switching module 200 and a beam optimization module 300;

[0049] The RFID tag module 100 constructs a three-dimensional reference coordinate system through the reference tag, and uses the time difference between the radio frequency signal reaching the reference tag and the target tag to calculate the three-dimensional position of the target tag in real time.

[0050] In order to better establish a three-dimensional coordinate reference system, the RFID tag module 100 fixes reference tags at four fixed points on the shelf where the target tag is located to construct a three-dimensional coordinate reference system. Each reference tag is equipped with a high-precision clock to support synchronous trigger signal transmission;

[0051] Four reference tags (denoted as R1, R2, R3, and R4) are deployed at four fixed points on the shelf where the target tag is located, forming non-coplanar points in three-dimensional space to ensure the uniqueness of the geometric solution. The reference tag Ri transmits a radio frequency pulse signal carrying an ID and a timestamp at time t. The signal propagation time is calculated and multipath reflection signals are eliminated by minimizing the signal arrival time (ToA).

[0052] In traditional distributed positioning systems, if each reference tag uses an independent clock, crystal oscillator frequency deviation can cause signal transmission time to desynchronize (e.g., microsecond offset), ultimately resulting in ranging errors (e.g., 1 microsecond at the speed of light corresponds to a 300-meter error). High-precision clocks (such as GPS-synchronized clocks, atomic clocks, or IEEE1588 / PTP protocol synchronization) can control the clock deviation of each reference tag to nanoseconds (e.g., ±10ns), corresponding to a ranging error of less than 3 cm.

[0053] In order to more accurately calculate the three-dimensional coordinates of the target tag, the RFID tag module 100 synchronizes the signal transmission time of each reference tag, measures the time difference between the radio frequency signal reaching the target tag, constructs a nonlinear equation system based on the time difference, and uses the least squares method to iteratively solve the three-dimensional coordinates of the target tag;

[0054] The constructed system of equations essentially describes the intersection of hyperboloids (in three-dimensional space, the locus of points with a constant distance difference to two fixed points is a hyperboloid). Multiple hyperboloids can be formed by using more than four reference labels, and their intersection is the target coordinate. The geometric meaning is clear and the solution is unique.

[0055] In actual scenarios, time measurement is subject to random noise (such as time errors caused by crystal oscillator jitter and multipath effects), which may result in large deviations when directly solving the system of equations. The least squares method evenly distributes the noise impact to each equation by minimizing the sum of squared errors, preventing a single outlier from dominating the solution. For nonlinear systems of equations (because the distance equation contains square terms), Taylor expansion is usually used for linearization and then iterative solution. The least squares method continuously approaches the optimal solution during the iterative process. Even if there are deviations in the initial value, it can converge to the global minimum through multiple rounds of iterations.

[0056] In addition, the direction switching module 200 obtains the three-dimensional position of the target tag, forms a phased array antenna through a power divider and a 3dB bridge, uses the power divider to distribute the input RF signal with equal amplitude, and combines the phase difference characteristics of the 3dB bridge to change the beam direction of the phased array to point to the position of the target tag.

[0057] like Figure 2 As shown, the direction switching module 200 distributes the input RF signal to the phased array with equal amplitude through the power divider, wherein the connection mode between the power divider and the bridge is:

[0058] The power splitter receives the radio frequency signal of the RFID reader, the door-entry output of the power splitter is connected to the input end of the bridge, and the exit output of the power splitter is connected to the isolation end of the bridge.

[0059] The radio frequency signal transmitted by the RFID reader is input to the power splitter, which distributes the input radio frequency signal with equal amplitude. Output 1 of the power splitter in the direction of entering the door is connected to the input end of bridge 1, output 2 of the power splitter in the direction of entering the door is connected to the input end of bridge 2, and output 3 of the power splitter in the direction of entering the door is connected to the input end of bridge 3. Output 1 of the power splitter in the direction of exiting the door is connected to the isolation end of bridge 1, output 2 of the power splitter in the direction of exiting the door is connected to the isolation end of bridge 2, and output 3 of the power splitter in the direction of exiting the door is connected to the isolation end of bridge 3.

[0060] The power divider outputs from the door entrance to the bridge input:

[0061] The signal distributed by the power divider enters the input end of the bridge, and after being processed by the bridge, it is output from the through end and the coupling end to form two signals with equal amplitude and orthogonal phase (90 degrees difference), which are used for phase modulation of the phased array;

[0062] The power splitter outputs to the isolation end of the bridge:

[0063] The other signal of the power divider is directly connected to the isolation end of the bridge. Ideally, there should be no signal at the isolation end, but in this design, the following functions may be achieved by actively injecting signals:

[0064] Cancel the reflected signal: By injecting the signal in reverse, the reflected wave inside the bridge is canceled, the standing wave ratio (VSWR) is reduced, and the signal loss is reduced;

[0065] Flexible adjustment of phase difference: In conjunction with the two signals of the power splitter, the bridge can generate multiple sets of phase differences (such as 0°, 90°, 180°, and 270°), providing more flexible phase control capabilities for the phased array and facilitating dynamic adjustment of the beam direction.

[0066] In order to better switch the beam direction and achieve circular polarization, the direction switching module 200 changes the excitation amplitude of the corresponding phased array by controlling the switching state of different output ports of the power divider, adjusts the phase difference between the two signals through a 3dB bridge, and switches the beam direction of the phased array;

[0067] The power divider distributes the input signal to different antenna unit channels. By controlling the attenuation or switching state of different output ports of the power divider, the excitation amplitude of the corresponding antenna unit can be changed. For example, for a 4-unit antenna array, a 1-to-4 power divider is used. When the beam needs to be directed to the left, the power divider is controlled so that the signal amplitude of the two units on the right is attenuated to zero, while the two units on the left are normally excited. In this way, the radiation beam of the antenna array will be deflected to the left. Conversely, when the beam needs to be directed to the right, a similar operation is performed on the left unit to achieve left-right switching of the beam.

[0068] The bridge can produce a specific phase difference. Using a 3dB bridge in conjunction with a power divider can further accurately adjust the phase between antenna units. A 3dB bridge can be used to split the input signal into two signals with a 90° phase difference. These two signals are distributed to different antenna unit groups through a power divider. If the beam needs to be switched to the left, the signal phase of the left antenna unit group can be advanced by a certain angle (such as 45°) than that of the right unit group. By adjusting the connection method and parameters of the bridge and power divider, precise pointing of the beam to the left can be achieved. When switching to the right is required, the direction of the phase difference is changed so that the signal phase of the right unit group is advanced, thereby achieving beam switching to the right;

[0069] like Figure 3 As shown, the phase difference characteristic of the 3dB bridge is used in combination with the power divider to adjust the phase of each antenna unit and change the beam direction of the array, for example:

[0070] Left beam: Element 1 (0°), Element 2 (90°), Element 3 (180°), Element 4 (270°) to linear phase gradient of +90° / element, beam deflected to the left;

[0071] Right beam: Element 1 (270°), Element 2 (180°), Element 3 (90°), Element 4 (0°) to linear phase gradient of -90° / element, beam deflected to the right;

[0072] The input signal is equally divided into multiple paths (such as two paths), which are fed into different 3dB bridges respectively to provide 0° and 90° phase signals for each group of antenna elements. The beam direction is switched by cascading bridges and adjusting the phase gradient.

[0073] like Figure 4 As shown in the figure, the RF input signal is divided into two paths by the bridge: one path is directly output (phase 0°), and the other path is output after being delayed by a 3dB bridge (phase 90°);

[0074] If the H polarization unit is connected to a 0° signal and the V polarization unit is connected to a 90° signal, right-hand circular polarization (RHCP) is synthesized.

[0075] If the H polarization unit is connected to a 90° signal and the V polarization unit is connected to a 0° signal, left-hand circular polarization (LHCP) is synthesized;

[0076] By switching the connection relationship between the bridge output port and the antenna unit, the switching between left and right circular polarization can be achieved.

[0077] In order to better reduce the electromagnetic coupling between phased arrays, the direction switching module 200 is provided with an electromagnetic bandgap structure and a high-impedance surface between adjacent arrays in the phased array antenna to suppress the mutual coupling between the arrays. At the same time, the isolation port of the 3dB bridge is connected to a 50-ohm matching resistor to absorb the reflected power.

[0078] The mutual coupling effect is more significant in wideband scenarios (for example, when the frequency offset is ±10% of the center frequency, the mutual coupling coefficient may increase by 5dB). The broadband suppression characteristics of EBG / HIS can broaden the effective operating bandwidth of the array and support ultra-wideband (UWB) phased array applications;

[0079] Mutual coupling can cause premature grating lobes (especially in arrays with large element spacing), limiting the beam scanning angle. Suppressing mutual coupling allows for larger element spacing, expanding the scanning range (e.g., from ±60° to ±80°) without grating lobes.

[0080] If the isolation port is left floating or has an impedance mismatch, high-power signals may flow back through the isolation port into the RF link (such as the power splitter and reader), causing low-noise amplifier (LNA) saturation or power amplifier (PA) overload. Matching resistors absorb power to provide overpower protection for front-end components. They are particularly suitable for high-power scenarios (such as RFID systems with a transmit power of ≥1W). Matching resistors can enhance the isolation of the bridge (for example, from 20dB to 25dB) and ensure that the amplitude difference between the two output signals is ≤0.5dB and the phase difference is strictly maintained at 90°. This is crucial for orthogonal phase modulation (such as IQ modulation) and polarization diversity in phased arrays.

[0081] In addition, the beam optimization module 300 collects environmental data and radio frequency signal data of the shelf where the target tag is located, uses the environmental data and radio frequency characteristic data as input, establishes a neural network model, outputs the optimal beam parameters pointing to the target tag, dynamically adjusts the pointing beam of the phased array, uses the tag recognition success rate as an indicator, evaluates the current beam performance, and performs feedback optimization on the beam.

[0082] In order to better construct the neural network model, the beam optimization module 300 generates RF propagation data under different environments through electromagnetic simulation software, records the corresponding optimal beam parameters, constructs a training set, and trains the neural network model;

[0083] Electromagnetic simulation software (such as HFSS and CST) can be used to simulate RF propagation characteristics in different scenarios (such as multiple obstacles indoors, metal reflective surfaces, and dynamic human movement). This generates data sets containing parameters such as path loss, multipath effects, and phase delay. Convolutional layers are used to extract spatial features such as environmental maps and obstacle distribution (for example, by using rasterized environmental modeling as image input). This captures the effects of geometric structures on signals (such as occlusion and diffraction). Fully connected layers are used to analyze the measured / simulated parameters of RF signals (such as power, phase, and angle of arrival) and establish a mapping relationship between electromagnetic propagation patterns and environmental structures.

[0084] By adjusting simulation parameters (such as frequency, antenna spacing, and obstacle material), massive virtual samples are generated to make up for the lack of measured data. This is especially suitable for extreme environments that are difficult to reproduce.

[0085] In order to better perform feature splicing, the beam optimization module 300 uses a convolutional layer to extract environmental image feature data, extracts radio frequency feature data through a fully connected layer, splices the environmental image feature data and the radio frequency feature data, and generates beam parameters through multiple layers of fully connected layers;

[0086] The model fuses environmental image features (spatial dimension) with RF features (electromagnetic dimension) through a stitching layer, breaking the limitations of traditional methods that separate environmental modeling and signal processing. For example, in a warehouse scenario, the model can simultaneously identify shelf layout (image features) and signal reflections caused by the shelves' metal surfaces (RF features), dynamically adjusting the beam to avoid strong reflections.

[0087] like Figure 5 As shown in the figure, two types of data, environment image and RF feature vector, are input in parallel. The convolutional layer of the convolutional neural network model (CNN) is used to extract features from the environment image. The RF feature vector is extracted through the fully connected layer. The 128-dimensional RF vector is input and the dimension is reduced by two layers of fully connected layers (128→64→32). The 32-dimensional feature vector is output. The visual features (128 dimensions) and the RF features (32 dimensions) are concatenated into a 160-dimensional fusion vector. The beam parameters are generated through multiple layers of fully connected layers (160→256→128):

[0088] Phase parameter: mapped to 0° to 360° through activation function;

[0089] Amplitude parameter: normalized to the range of 0 to 1;

[0090] Polarization mode: The classifier outputs discrete values ​​(such as 0 / 1 for left-handed and right-handed circular polarization).

[0091] In order to better calculate the tag recognition success rate, the beam optimization module 300 records the number of responses of each tag through the RFID reader, calculates the proportion of tags that successfully respond, and performs multiple repeated tests to take the average value to eliminate random errors;

[0092] Under a fixed shelf layout, place tags with a known number and location (e.g., 100 tags) to simulate a shopping scenario. Allow the tags (e.g., attached to merchandise) to move at different speeds through the beam coverage area. Record the number of responses from each tag using an RFID reader, and calculate the percentage of tags that successfully respond. Repeat the test more than 10 times and take the average value to eliminate random errors.

[0093] In order to better optimize the beam, the beam optimization module 300 divides the tag recognition success rate into four evaluation levels: excellent, good, average, and poor, and formulates corresponding optimization suggestions;

[0094] Tag recognition success rate evaluation:

[0095]

[0096] If the recognition rate on the left side of the shelf is low, the beam is deflected 15° to the left through the phased array control algorithm. When the linear polarization TISR is <85%, it is switched to circular polarization (such as RHCP). The adaptability of circular polarization to tags of any orientation is used to improve the recognition rate. The power weight of the antenna unit in the weak coverage area is increased (such as +3dB) to improve the local signal strength.

[0097] In summary, the working principle of this solution is as follows:

[0098] This RFID high-gain phased array antenna array uses the RFID tag module 100 to establish a coordinate system. It uses the time difference between the RF signal reaching the reference tag and the target tag to calculate the three-dimensional position of the target tag in real time. The direction switching module 200 obtains the three-dimensional position of the target tag. The phased array antenna is formed by a power divider and a 3dB bridge. The power divider is used to distribute the input RF signal with equal amplitude. Combined with the phase difference characteristics of the 3dB bridge, the signal phase difference is adjusted by the 3dB bridge, achieving circular polarization radiation of the antenna and flexible switching of the beam direction. Circular polarization enhances the antenna's ability to receive tags in different directions. The beam direction switching function can adjust the signal coverage area according to the needs of retail scenarios and improve the tag recognition rate.

[0099] The beam optimization module 300 uses a neural network model to calculate the optimal beam parameters of the target tag based on the environmental data and radio frequency signal data of the shelf where the target tag is located, and dynamically adjusts the pointing beam of the phased array to achieve dynamic adjustment of the beam signal parameters according to the environment near the target tag, reducing multipath interference. At the same time, the tag recognition success rate is used as an indicator to evaluate the current beam performance and perform feedback optimization on the beam.

[0100] The above shows and describes the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions are merely preferred examples of the present invention and are not intended to limit the present invention. Various changes and improvements may be made to the present invention without departing from the spirit and scope of the present invention. Such changes and improvements fall within the scope of the present invention. The scope of protection claimed in the present invention is defined by the appended claims and their equivalents.

Claims

1. RFID high-gain phased array antenna array, characterized by: It comprises an RFID tag module (100), a direction switching module (200) and a beam optimization module (300); The RFID tag module (100) constructs a three-dimensional reference coordinate system through a reference tag, and uses the time difference between the radio frequency signal reaching the reference tag and the target tag to calculate the three-dimensional position of the target tag in real time; The direction switching module (200) obtains the three-dimensional position of the target tag, forms a phased array antenna through a power divider and a 3dB bridge, uses the power divider to distribute the input radio frequency signal in equal amplitude, and combines the phase difference characteristics of the 3dB bridge to change the beam direction of the phased array to point to the position of the target tag; The beam optimization module (300) collects environmental data and radio frequency signal data of the shelf where the target tag is located, takes the environmental data and radio frequency characteristic data as input, establishes a neural network model, outputs optimal beam parameters pointing to the target tag, dynamically adjusts the pointing beam of the phased array, evaluates the current beam performance with the tag recognition success rate as an indicator, and performs feedback optimization on the beam.

2. The RFID high-gain phased array antenna array according to claim 1, characterized in that: The RFID tag module (100) fixes reference tags at four fixed points on the shelf where the target tag is located to construct a three-dimensional coordinate reference system. Each reference tag is equipped with a high-precision clock to support synchronous trigger signal transmission.

3. The RFID high-gain phased array antenna array according to claim 2, characterized in that: The RFID tag module (100) synchronizes the signal transmission time of each reference tag, measures the time difference between the radio frequency signal and the target tag, constructs a nonlinear equation group using the time difference, and adopts the least square method to iteratively solve the three-dimensional coordinates of the target tag.

4. The RFID high-gain phased array antenna array according to claim 1, characterized in that: The direction switching module (200) distributes the input radio frequency signal to the phased array with equal amplitude through a power divider, wherein the power divider is connected to the bridge in the following manner: The power splitter receives the radio frequency signal of the RFID reader, the door-entry output of the power splitter is connected to the input end of the bridge, and the exit output of the power splitter is connected to the isolation end of the bridge.

5. The RFID high-gain phased array antenna array according to claim 1, characterized in that: The direction switching module (200) changes the excitation amplitude of the corresponding phased array by controlling the switch states of different output ports of the power divider, adjusts the phase difference between two signals through a 3dB bridge, and switches the beam direction of the phased array.

6. The RFID high-gain phased array antenna array according to claim 5, characterized in that: The direction switching module (200) is provided in a phased array antenna, with an electromagnetic bandgap structure and a high impedance surface provided between adjacent arrays to suppress mutual coupling between the arrays; and at the same time, a 50-ohm matching resistor is connected to the isolation port of the 3dB bridge to absorb reflected power.

7. The RFID high-gain phased array antenna array according to claim 1, characterized in that: The beam optimization module (300) generates radio frequency propagation data under different environments through electromagnetic simulation software, records corresponding optimal beam parameters, constructs a training set, and trains a neural network model.

8. The RFID high-gain phased array antenna array according to claim 7, characterized in that: The beam optimization module (300) extracts environmental image feature data using a convolutional layer, extracts radio frequency feature data using a fully connected layer, splices the environmental image feature data and the radio frequency feature data, and generates beam parameters using multiple layers of fully connected layers.

9. The RFID high-gain phased array antenna array according to claim 1, characterized in that: The beam optimization module (300) records the number of responses of each tag through an RFID reader, counts the proportion of tags that respond successfully, and performs multiple repeated tests to take an average value to eliminate random errors.

10. The RFID high-gain phased array antenna array according to claim 9, characterized in that: The beam optimization module (300) divides the tag recognition success rate into four evaluation levels: excellent, good, average and poor, and formulates corresponding optimization suggestions.

Images