Low-power RFID group scanning system based on energy harvesting and backscattering technology

Abstract

The present invention provides a low-power RFID group scanning system based on energy collection and backscattering technology, which relates to the technical field of Internet of Things; the RFID tag in the system integrates a radio frequency receiving component, a photovoltaic component and a thermoelectric component, and is mainly used to collect radio frequency energy in the environment, and is supplemented by other energy sources, and energy conversion and storage are performed through a maximum power point tracking algorithm to ensure the stability of tag power supply; the RFID reader uses undersampling and sparse reconstruction technology to parse backscattered signals, and supports efficient identification and information extraction of multiple tags; the data management unit is responsible for storing and managing RFID tag information, and establishing a logical association between an optical fiber port and a pigtail; the system has efficient energy utilization capability, intelligent signal processing and a highly adaptable dynamic adjustment mechanism, and is widely applicable to optical fiber network management and low-power Internet of Things application scenarios.

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to a low-power RFID group scanning system based on energy collection and backscattering technology. Background Art

[0002] With the rapid development of Internet of Things (IoT) technology, RFID (Radio Frequency Identification) technology is increasingly used in various fields. Especially in scenarios such as logistics management, equipment monitoring and industrial automation, RFID technology has gradually become an important solution due to its non-contact, long-distance and automatic identification characteristics. However, since traditional RFID tags usually rely on external batteries for power supply, their service life, maintenance costs and environmental adaptability are subject to certain limitations. Therefore, low-power RFID technology based on energy harvesting has emerged and has become a research direction of concern in the industry.

[0003] To this end, for example, a Chinese patent application with authorization announcement number CN110770754B discloses an energy harvesting RFID circuit, an energy harvesting RFID tag, and related methods. The RFID circuit receives an interrogation signal from an antenna on an input lead and collects energy, which is stored in a charge storage unit. The boost converter then boosts the collected voltage to a boosted voltage, which is fed to the sensor output lead as an output voltage. The patent also provides a booster switch unit for activating the boost converter until the voltage output by the charge storage unit reaches a first threshold, and the booster switch unit activates the boost converter. One or more sensor input leads receive a measurement value input based on the output voltage, and the measurement value input reflects the sensed value of the object to be measured, such as strain in a pipe.

[0004] However, the above methods still have many shortcomings in terms of energy collection efficiency, response speed, environmental adaptability and signal processing capabilities, and it is difficult to meet the requirements of efficient and reliable operation of low-power RFID systems in complex application scenarios. Therefore, there is an urgent need for a low-power RFID system that can simultaneously achieve multi-energy collection, dynamic energy management, and efficient signal processing to solve the above problems and meet the needs of practical applications. Summary of the invention

[0005] In view of the deficiencies of the prior art, the present invention provides a low-power RFID group scanning system based on energy collection and backscattering technology.

[0006] The present invention provides a low-power RFID group scanning system based on energy collection and backscattering technology, comprising:

[0007] A plurality of RFID tags are attached to the optical fiber port and the pigtail respectively; wherein each of the RFID tags comprises: a collection unit, used to collect radio frequency energy from the environment and convert it into electrical energy; a conversion and storage unit, based on the maximum power point tracking method, using a preset energy storage element to convert and store the electrical energy and power the RFID tag; a scattering communication unit, including an antenna and a signal reflection unit, wherein the antenna is used to receive a radio frequency excitation signal from an RFID reader, and the signal reflection unit is used to reflect a modulated signal containing tag information back to the RFID reader by backscattering;

[0008] The RFID reader / writer includes: a radio frequency transmitting unit, used to send a radio frequency excitation signal to the RFID tag; a radio frequency receiving unit, used to receive a backscattered signal from the RFID tag; an intelligent signal processing unit, connected to the radio frequency receiving unit, and based on a multi-path signal processing method, analyzing the received backscattered signal and distinguishing signals from multiple RFID tags;

[0009] The data management unit is used to collect, store and manage the read RFID tag information and automatically establish a logical association between the optical fiber port label and the pigtail label.

[0010] As an optional implementation, the collection unit is also used to collect light energy and heat energy from the environment;

[0011] The radio frequency energy is collected from the radio wave energy in the environment through the radio frequency receiving component and converted into electrical energy;

[0012] Collect light energy through photovoltaic modules and convert it into electrical energy;

[0013] Collect ambient temperature difference energy through thermoelectric components and convert it into electrical energy;

[0014] The electric energy collected by the radio frequency receiving component, the photovoltaic component and the thermoelectric component is input into the conversion and storage unit to perform electric energy integration and output control.

[0015] As an optional implementation manner, the maximum power point tracking method includes:

[0016] Collect environmental data, combine it with historical communication event data, use the sliding window method to calculate the probability and time interval of communication events, and generate prediction results;

[0017] Based on the prediction result, determining the triggering time and energy requirement of the communication event, and generating an energy collection path adjustment instruction;

[0018] The energy collection path adjustment instruction includes adjusting input parameters of each energy collection component and triggering conditions for mode switching; the energy collection component includes a radio frequency receiving component, a photovoltaic component and a thermoelectric component;

[0019] In response to the energy harvesting path adjustment instruction not triggering the switching condition, the conversion and storage unit enters a low-power sleep mode and only maintains low-frequency monitoring of environmental data;

[0020] In response to the energy collection path adjustment instruction triggering a switching condition and a communication event about to occur, activating the main energy collection path, and adjusting input parameters of the radio frequency receiving component, the photovoltaic component, and the thermoelectric component;

[0021] In response to the actual triggering of the communication event, the conversion and storage unit supplies power to the scattering communication unit at a first priority according to the allocation scheme determined by the energy collection path adjustment instruction;

[0022] In response to the completion of the communication event, the default parameter configuration is restored based on the energy collection path adjustment instruction, the low power sleep mode is switched back, and the environment monitoring state is re-entered.

[0023] As an optional implementation, the sliding window method includes:

[0024] Set sliding windows to update and manage collected environmental data and historical communication event data in real time;

[0025] In each sliding window, the number of communication events and the time interval are counted, and the average occurrence rate of communication events is calculated;

[0026] An improved probability model is used to calculate the occurrence probability and predicted time interval of a communication event according to statistical results, wherein the probability model is a piecewise Poisson process or a two-state Markov chain model;

[0027] Based on the probability of communication event occurrence and the predicted time interval, a prediction result for energy harvesting path adjustment is generated.

[0028] As an optional implementation, the improved probability model includes:

[0029] Divide the time into a plurality of preset time periods, consider the occurrence rate of communication events in each time period to be constant, and calculate the Poisson parameter of each time period;

[0030] For the two states of communication event occurrence and non-occurrence, the probability of communication event occurrence at the next moment is predicted based on the state and state transition probability at the previous moment.

[0031] As an optional implementation, the length of the sliding window and the sliding step are adaptively adjusted according to the frequency of communication events and the rate of environmental change.

[0032] As an optional implementation manner, the generating the energy collection path adjustment instruction includes:

[0033] Calculate the activation function E(t), where E(t) reflects the energy state and communication requirements of the low-power RFID group scanning system at time t;

[0034] When the excitation function E(t) reaches the preset trigger condition, an energy collection path adjustment instruction is generated to adjust the input parameters of each energy collection component and switch the energy collection mode;

[0035] When the excitation function E(t) does not reach the trigger condition, the current energy collection mode is maintained, and the conversion and storage unit remains in a low-power sleep mode.

[0036] As an optional implementation, the calculation of the activation function E(t) includes:

[0037] ;

[0038] in is the predicted probability of the communication event occurring at time t, is the energy storage at time t, is the maximum capacity of energy storage, is the energy consumption rate at time t, , , The weight parameters are set according to the requirements of the low-power RFID group scanning system.

[0039] As an optional implementation manner, the multipath signal processing method includes:

[0040] Under-sampling the received backscattered signal to obtain compressed sampling data of the signal;

[0041] Using sparse reconstruction algorithm, the original signal is reconstructed based on the compressed sampling data;

[0042] Demodulate and analyze the reconstructed original signal to extract the data information of each RFID tag;

[0043] Wherein, the sparse reconstruction algorithm includes:

[0044] Based on a preset sparse basis function, the compressed sampling data is mapped to a sparse domain to establish a sparse representation model of the signal; wherein the sparse basis function is a discrete cosine transform basis;

[0045] Based on the iterative hard threshold method, the sparse representation model is solved to reconstruct the original signal. The iterative hard threshold method performs the following steps in each iteration:

[0046] Calculate the gradient of the current iteration and obtain the update direction of the signal;

[0047] Perform a hard threshold operation on the gradient update result, retaining only the largest number of coefficients specified by the sparsity parameter, and setting the rest of the coefficients to zero;

[0048] Using the coefficients processed by hard threshold, an updated signal estimate is obtained through inverse transformation, and this is used as a reference for the reconstructed signal in the next iteration;

[0049] Repeat the above iterative process until the preset convergence condition is met or the maximum number of iterations is reached.

[0050] Compared with the prior art, the beneficial effects of the present invention are as follows: the present invention realizes efficient energy collection and dynamic management by integrating multiple energy collection methods such as radio frequency, photovoltaic, and thermoelectric, and combining the maximum power point tracking algorithm; through the sliding window method and the improved probability model, the system can intelligently predict communication events and adjust the energy collection path in advance to ensure the reliability of energy supply; the use of undersampling and sparse reconstruction technology effectively solves the problem of signal separation and analysis in a multi-tag environment, and improves recognition efficiency and accuracy. At the same time, the present invention has the characteristics of strong dynamic adaptability, low power consumption, and fast intelligent response, which can extend the service life of RFID tags and simplify the management process of optical fiber networks. It is particularly suitable for complex scenarios such as data centers and communication rooms, and fully meets the needs of modern Internet of Things applications for efficient, intelligent, and low-power RFID systems. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] Figure 1 A schematic diagram of a low-power RFID group scanning system based on energy harvesting and backscattering technology provided by an embodiment of the present invention;

[0052] Figure 2 A schematic diagram of a collection unit provided by an embodiment of the present invention;

[0053] Figure 3 A flow chart of a multipath signal processing method provided for the implementation of the present invention.

[0054] Reference numerals:

[0055] 10. RFID tag; 11. Collection unit; 111. Photovoltaic module; 112. Thermoelectric module; 113. Radio frequency receiving module; 12. Conversion and storage unit; 13. Scattering communication unit; 131. Antenna; 132. Signal reflection unit;

[0056] 20. RFID reader / writer; 21. Radio frequency transmitting unit; 22. Radio frequency receiving unit; 23. Intelligent signal processing unit; 30. Data management unit. DETAILED DESCRIPTION

[0057] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0058] See also Figure 1 , Figure 1 A schematic diagram of a low-power RFID group scanning system based on energy harvesting and backscattering technology provided by an embodiment of the present invention. The system includes multiple RFID tags 10, an RFID reader 20, and a data management unit 30;

[0059] The multiple RFID tags 10 are attached to the optical fiber port and the pigtail respectively; wherein each of the RFID tags 10 comprises:

[0060] The collecting unit 11 is used to collect radio frequency energy from the environment and convert it into electrical energy; the collecting unit 11 includes a photovoltaic component 111, a thermoelectric component 112 and a radio frequency receiving component 113;

[0061] A conversion and storage unit 12, connected to the collection unit 11, converts and stores the electric energy using a preset energy storage element based on a maximum power point tracking method, and supplies power to the RFID tag 10;

[0062] The scattering communication unit 13 includes an antenna 131 and a signal reflecting unit 132, wherein the antenna 131 is used to receive a radio frequency excitation signal from the RFID reader 20, and the signal reflecting unit 132 is used to reflect a modulated signal containing tag information back to the RFID reader 20 by backscattering;

[0063] The RFID reader 20 includes:

[0064] A radio frequency transmitting unit 21, used to send a radio frequency excitation signal to the RFID tag 10;

[0065] A radio frequency receiving unit 22, used to receive a backscattered signal from the RFID tag 10;

[0066] An intelligent signal processing unit 23, connected to the radio frequency receiving unit 22, analyzes the received backscattered signal based on a multipath signal processing method, and distinguishes signals from a plurality of the RFID tags 10;

[0067] The data management unit 30 communicates with the intelligent signal processing unit 23 and is used to collect, store and manage the read RFID tag information and automatically establish a logical association between the optical fiber port label and the pigtail label.

[0068] In a specific implementation, a plurality of RFID tags 10 are attached to the optical fiber port and the pigtail respectively, and each RFID tag 10 integrates a collection unit 11, a conversion and storage unit 12, and a scattering communication unit 13. The collection unit 11 is used to collect light energy, heat energy, and radio frequency energy from the environment and convert them into electrical energy. Among them, light energy is collected by a micro photovoltaic cell integrated on the surface of the tag, and the light energy is converted into electrical energy using ambient light (including natural light and indoor light); thermal energy is collected by a micro thermoelectric converter installed at the contact position between the tag and the surface of the device, and the temperature difference generated by the operation of the device is used to convert the thermal energy into electrical energy; radio frequency energy is collected from radio waves (such as Wi-Fi and Bluetooth signals) in the environment through a broadband antenna and a rectifier circuit, and the AC radio frequency signal is converted into DC power.

[0069] The conversion and storage unit 12 is connected to the collection unit 11, and based on the maximum power point tracking (MPPT) algorithm, it monitors the output voltage and current of the photovoltaic module 111 and the thermoelectric module 112 in real time, and achieves the maximum power output by adjusting the load. The unit uses a high-energy-density supercapacitor or lithium-ion micro-battery as an energy storage element to store the electric energy obtained from the collection unit 11 and provide a stable power supply for the scattering communication unit 13. The energy management circuit includes a power management IC, which is responsible for integrating the electric energy from different energy sources, performing voltage regulation, overcharge and over-discharge protection, etc., to ensure the safety of energy storage elements and electrical equipment.

[0070] The antenna 131 uses a miniaturized patch antenna or a helical antenna to receive the excitation signal of the RFID reader 20 and send the backscattered signal. The core of the signal reflection unit 132 is a backscatter modulator, which is usually composed of a field effect transistor (FET). The control chip (such as a low-power microcontroller) inside the RFID tag 10 controls the switching state of the FET according to the stored tag information, modulates the reflection coefficient of the antenna 131, and reflects the modulated signal containing the tag information back to the RFID reader 20.

[0071] The RF transmitting unit 21 sends a RF excitation signal of a specific frequency (such as 915 MHz) to space through a transmitting antenna, and the power and frequency of the excitation signal meet the requirements of relevant regulations. The RF receiving unit 22 receives the backscattered signal from the RFID tag 10 through a receiving antenna. The signal is amplified by a low noise amplifier (LNA), the filter filters out noise, and the analog-to-digital converter (ADC) converts the analog signal into a digital signal. The intelligent signal processing unit 23 is connected to the RF receiving unit 22, and analyzes the received backscattered signal based on a multipath signal processing method, and distinguishes the signals from multiple RFID tags 10. Due to the multipath effect in the environment, the received signal may be delayed and attenuated. The intelligent signal processing unit 23 can use a preset algorithm (such as fast Fourier transform, matched filter) to process the signal and extract useful information; use multi-user detection technology (such as code division multiple access or orthogonal frequency division multiple access) to distinguish the signals from multiple RFID tags 10 and avoid signal conflicts.

[0072] The data management unit 30 communicates with the intelligent signal processing unit 23, and is used to collect, store and manage the read RFID tag 10 information, and automatically establish a logical association between the fiber port label and the pigtail label. The unit establishes a database to record the unique identification information of each RFID tag 10, as well as the corresponding fiber port or pigtail position. By reading the information of the fiber port label and the pigtail label, the connection relationship between them is automatically established, which is convenient for subsequent network topology management. The data management unit 30 provides a user interface to support real-time monitoring, query and management of the fiber network connection status.

[0073] Exemplarily, the collection unit 11 of the RFID tag 10 continuously collects light energy, heat energy and radio frequency energy from the environment. The conversion and storage unit 12 optimizes the energy collection efficiency through the MPPT algorithm and stores the electrical energy in the energy storage element. In the non-communication state, the RFID tag 10 is in a low-power sleep mode to reduce energy consumption and only maintain the basic energy collection function. When the RFID reader 20 sends a radio frequency excitation signal, the scattering communication unit 13 of the RFID tag 10 is activated and enters the working mode. After the RFID tag 10 receives the excitation signal, the scattering communication unit 13 reflects the modulated signal containing the tag information back to the RFID reader 20 through backscattering modulation. The radio frequency receiving unit 22 of the RFID reader 20 receives the backscattered signal, and the intelligent signal processing unit 23 parses the signal to extract the information of each RFID tag 10. The RFID reader 20 sends the acquired tag information to the data management unit 30, and the data management unit 30 updates the database, maintains the logical association between the optical fiber port and the pigtail, and realizes the automatic management of the optical fiber network.

[0074] The present invention has important application value in large data centers or communication rooms. In these scenarios, the number of optical fiber connections is huge, and the traditional manual recording and management methods are inefficient and prone to errors. After adopting the present invention, the RFID tag 10 is pasted or embedded on the optical fiber port and the pigtail, and no external power supply or battery replacement is required, which is easy to install. Through the periodic scanning of the RFID reader 20, the changes in the optical fiber connection status are obtained in real time, and abnormalities are discovered in time. The data management unit 30 can be connected to the upper-level network management system to achieve unified management of the physical layer and the logical layer, and improve maintenance efficiency.

[0075] The advantages of this system are its self-powered design, which uses environmental energy to power the RFID tag 10, reducing maintenance costs and improving system reliability; high-concurrency recognition, using advanced signal processing technology, can simultaneously identify a large number of RFID tags, and is suitable for complex network environments; low-power operation, the RFID tag 10 enters sleep mode when not in operation, greatly extending the service life of the device.

[0076] See also Figure 2 , Figure 2 A schematic diagram of a collection unit provided by an embodiment of the present invention;

[0077] As an optional implementation, the collection unit 11 is also used to collect light energy and heat energy from the environment;

[0078] The radio frequency receiving component 113 collects radio frequency energy from the radio wave energy in the environment and converts it into electrical energy;

[0079] Collecting light energy through photovoltaic modules 111 and converting it into electrical energy;

[0080] The ambient temperature difference energy is collected by the thermoelectric component 112 and converted into electrical energy;

[0081] The electric energy collected by the RF receiving component 113 , the photovoltaic component 111 and the thermoelectric component 112 is input into the conversion and storage unit 12 for electric energy integration and output control.

[0082] In a specific implementation, first, light energy is collected and converted into electrical energy through the photovoltaic module 111. For example, the photovoltaic module 111 can use a high-efficiency micro photovoltaic cell, which can be a single crystal silicon, polycrystalline silicon or thin film solar cell, installed on the surface or side of the RFID tag 10 to maximize the light receiving area. Even indoors or in a low-light environment, the photovoltaic module 111 can use artificial lighting or scattered light to collect energy. In order to improve the energy conversion efficiency, an anti-reflection coating can be added to the surface of the photovoltaic module 111 to reduce light reflection losses. At the same time, the series-parallel structure of the battery is optimized to meet the voltage and current requirements of the RFID tag 10.

[0083] Secondly, the ambient temperature difference energy is collected and converted into electrical energy through the thermoelectric component 112. The thermoelectric component 112 uses the Seebeck effect to directly convert the temperature difference into voltage. The thermoelectric component 112 of the RFID tag 10 is installed on its back and is in close contact with the surface of the attached device. The thermoelectric component 112 can continuously generate electrical energy by using the temperature difference between the heat generated when the device is running and the ambient temperature. The thermoelectric material can be selected from high-performance bismuth telluride (Bi2Te3) alloy, which has a high thermoelectric conversion efficiency. In order to enhance the temperature difference, the RFID tag 10 structure can be designed as a double layer, adding thermal insulation material to ensure the temperature difference between the hot end and the cold end.

[0084] Again, RF energy is collected from the radio wave energy in the environment and converted into electrical energy through the RF receiving component 113. The RF receiving component 113 includes a broadband receiving antenna and a rectifier circuit. The receiving antenna can be designed as a spiral antenna, a dipole antenna or a patch antenna to adapt to environmental RF signals of different frequencies, such as Wi-Fi (2.4 GHz), cellular networks (900MHz, 1800 MHz) and TV broadcast signals. The RF signal received by the broadband receiving antenna passes through a rectifier circuit, usually using a high-frequency Schottky diode or a synchronous rectifier, to convert the AC signal into a DC voltage. In order to increase the voltage level, a multi-stage voltage doubling circuit can be used to increase the output voltage to meet the working requirements of the RFID tag 10.

[0085] Finally, the collected electric energy is input into the conversion and storage unit 12 for electric energy integration and output control. The conversion and storage unit 12 includes an energy management circuit and an energy storage element. The energy management circuit is responsible for integrating the electric energy from the photovoltaic component 111, the thermoelectric component 112 and the radio frequency receiving component 113, and uses a diode or a power switch to prevent reverse current between the energy sources. The energy management circuit also includes a boost or buck converter to adjust the input electric energy of different voltage levels to a uniform voltage to power the energy storage element and the load. The energy storage element can use a high-energy-density micro lithium-ion battery or a supercapacitor to provide a stable power output.

[0086] It should be noted that the energy collection system of the RFID tag 10 needs to be optimized according to the environmental conditions. For example, in an environment with sufficient light, the energy collection capacity of the photovoltaic component 111 is enhanced; in places with obvious temperature differences, such as the heat dissipation part of the equipment, the design of the thermoelectric component 112 is strengthened; in areas with rich radio frequency signals, such as near communication base stations, the performance of the radio frequency receiving component 113 is optimized. In addition, in order to achieve the optimization of energy collection, the conversion and storage unit 12 can adopt maximum power point tracking (MPPT) technology. The MPPT algorithm monitors the output characteristics of each energy collection component in real time, and adjusts the circuit parameters so that each component always works at the maximum power output point. Commonly used MPPT algorithms include perturbation observation method and incremental conductance method, and the algorithm is implemented by a microcontroller or a dedicated integrated circuit. The energy collection component includes a radio frequency receiving component 113, a photovoltaic component 111 and a thermoelectric component 112.

[0087] During the energy integration process, the conversion and storage unit 12 will intelligently allocate the input of each energy source according to the power state of the energy storage element and the power consumption requirements of the scattering communication unit 13. If the energy storage element has sufficient power, the system can enter a low-power mode to reduce energy collection activities and extend the life of the element. If it is detected that the energy storage power is lower than the set threshold, the system will give priority to opening the most efficient energy collection path to quickly replenish power.

[0088] In this way, the RFID tag 10 can continuously and stably obtain energy from the environment to provide power support for its own operation without replacing batteries or external power supplies, meeting the needs of low-power RFID group scanning systems. This multi-energy source collection method improves the reliability and adaptability of the system, allowing the RFID tag 10 to work normally under various environmental conditions.

[0089] As an optional implementation manner, the maximum power point tracking method includes:

[0090] Collect environmental data, combine it with historical communication event data, use the sliding window method to calculate the probability and time interval of communication events, and generate prediction results;

[0091] Based on the prediction result, determining the triggering time and energy requirement of the communication event, and generating an energy collection path adjustment instruction;

[0092] The energy collection path adjustment instruction includes adjusting the input parameters of each energy collection component and the triggering condition of the mode switching;

[0093] In response to the energy collection path adjustment instruction not triggering the switching condition, the conversion and storage unit 12 enters a low-power sleep mode and only maintains low-frequency monitoring of environmental data;

[0094] In response to the energy collection path adjustment instruction triggering the switching condition and the communication event is about to occur, activating the main energy collection path, and adjusting the input parameters of the radio frequency receiving component 113, the photovoltaic component 111 and the thermoelectric component 112;

[0095] In response to the actual triggering of the communication event, the conversion and storage unit 12 supplies power to the scattering communication unit 13 with a first priority according to the allocation scheme determined by the energy collection path adjustment instruction;

[0096] In response to the completion of the communication event, the default parameter configuration is restored based on the energy collection path adjustment instruction, the low power sleep mode is switched back, and the environment monitoring state is re-entered.

[0097] In this invention, in order to improve the efficiency of energy collection and utilization, an improved maximum power point tracking (MPPT) method is proposed. This method collects environmental data and historical communication event data, combines the sliding window algorithm, predicts the probability and time interval of communication events, and dynamically adjusts the energy collection path and energy storage strategy.

[0098] In the specific implementation, first, the sensor module built into the RFID tag 10 continuously collects environmental data, including light intensity, temperature change, RF signal strength, etc. These data reflect the energy collection capabilities of the photovoltaic component 111, the thermoelectric component 112, and the RF receiving component 113. At the same time, the system records the data of historical communication events, such as the occurrence time, duration, and interval time of the communication event. These data are stored in the memory of the RFID tag 10 for subsequent analysis.

[0099] Next, the above data is processed using the sliding window method. Specifically, a sliding window can be set, and the window length and sliding step size can be adaptively adjusted according to the frequency of communication events and the rate of environmental change. For example, for high-frequency communication events, a shorter window length and step size can be used to improve the timeliness of the prediction. In each sliding window, the number of occurrences and time intervals of communication events are counted, and the average occurrence rate of communication events is calculated.

[0100] Then, the probability of occurrence of communication events is calculated using an improved probability model, such as a segmented Poisson process and a two-state Markov chain model. For the segmented Poisson process, time is divided into multiple preset time periods (such as day and night, weekdays and weekends), and the Poisson parameters of each time period are calculated assuming that the occurrence rate of communication events in each time period is constant. For the two-state Markov chain model, the state of the communication event is divided into occurrence and non-occurrence, and the probability of the communication event occurring at the next moment is predicted based on the state at the previous moment and the known state transition probability.

[0101] Based on the above prediction results, the possible triggering time of the communication event and the corresponding energy demand are determined. Then, an energy collection path adjustment instruction is generated, which includes adjusting the input parameters of each energy collection component and the triggering condition of the mode switching. For example, the operating voltage of the photovoltaic component 111 is adjusted to optimize its output power, the load of the thermoelectric component 112 is adjusted to match the temperature difference change, or the resonant frequency of the RF receiving component 113 is adjusted to adapt to the spectrum change of the ambient RF signal.

[0102] When the energy collection path adjustment instruction does not trigger the switching condition, the conversion and storage unit 12 will enter the low-power sleep mode. At this time, the system only maintains low-frequency monitoring of environmental data, and the energy collection component works in the lowest power consumption state to reduce energy consumption and extend the service life of the energy storage element.

[0103] When the energy collection path adjustment instruction triggers the switching condition and the communication event is predicted to occur, the system will activate the main energy collection path. The conversion and storage unit 12 adjusts the input parameters of each energy collection component according to the instruction so that it works in the best state to quickly accumulate enough energy to meet the communication needs. For example, increase the MPPT sampling frequency of the photovoltaic component 111, increase the load current of the thermoelectric component 112, or enhance the sensitivity of the RF receiving component 113.

[0104] When a communication event is actually triggered, the conversion and storage unit 12 supplies power to the scattering communication unit 13 with the first priority according to the predetermined energy allocation scheme. At this time, other non-critical modules may temporarily reduce power consumption or enter a dormant state to ensure that the scattering communication unit 13 has enough energy for data transmission.

[0105] After the communication event is completed, the system will adjust the instructions based on the energy collection path, restore the default parameter configuration, switch back to low-power sleep mode, and re-enter the environmental monitoring state. In this way, the system can meet the communication needs while maximizing energy conservation and extending the working life of the equipment.

[0106] It should be noted that in the specific implementation process, the microcontroller or energy management chip in the system is responsible for executing the above algorithms and control logic. In order to achieve low power consumption and high efficiency, the microcontroller should have the following characteristics:

[0107] Low power design: Use low power mode and sleep mode to reduce energy consumption.

[0108] Fast wake-up capability: Ability to quickly switch from sleep mode to working mode before a communication event occurs.

[0109] High-precision ADC: used to collect environmental data and operating parameters of energy harvesting components.

[0110] Programmability: supports adjustment of algorithm parameters and control strategies according to actual application requirements.

[0111] In addition, the software part of the system needs to have functional modules such as data acquisition, data storage, algorithm calculation and instruction execution.

[0112] In this way, the system can intelligently predict the occurrence of communication events and adjust the energy collection strategy in advance to ensure that there is enough energy to supply the scattering communication unit 13 when needed. At the same time, in the non-communication state, the system maintains low power consumption operation to extend the service life of the device. This dynamic energy management method improves the reliability and adaptability of the system and meets the needs of low-power RFID group scanning systems.

[0113] As an optional implementation, the sliding window method includes:

[0114] Set sliding windows to update and manage collected environmental data and historical communication event data in real time;

[0115] In each sliding window, the number of communication events and the time interval are counted, and the average occurrence rate of communication events is calculated;

[0116] An improved probability model is used to calculate the occurrence probability and predicted time interval of a communication event according to statistical results, wherein the probability model is a piecewise Poisson process or a two-state Markov chain model;

[0117] Based on the probability of communication event occurrence and the predicted time interval, a prediction result for energy harvesting path adjustment is generated.

[0118] As an optional implementation, the improved probability model includes:

[0119] Divide the time into a plurality of preset time periods, consider the occurrence rate of communication events in each time period to be constant, and calculate the Poisson parameter of each time period;

[0120] For the two states of communication event occurrence and non-occurrence, the probability of communication event occurrence at the next moment is predicted based on the state and state transition probability at the previous moment.

[0121] As an optional implementation, the length of the sliding window and the sliding step are adaptively adjusted according to the frequency of communication events and the rate of environmental change.

[0122] In the present invention, in order to accurately predict the occurrence of communication events and thus optimize the adjustment of energy collection paths, a sliding window method is used to update and manage the collected environmental data and historical communication event data in real time.

[0123] In the specific implementation, first, set the sliding window to update and manage the collected environmental data and historical communication event data in real time. The length of the sliding window (window size) and the sliding step can be adaptively adjusted according to the frequency of communication events and the rate of environmental change. In general, the length of the sliding window can be set to an integer multiple of the average interval time of communication events to ensure that the window contains enough event data and improve the reliability of statistics. The sliding step determines the speed of the window movement and is usually set to a fixed proportion of the window length, such as 50%, to balance the computational complexity and real-time performance.

[0124] In the specific implementation, first, the time is divided into multiple preset time periods, the occurrence rate of communication events in each time period is considered constant, and the Poisson parameters of each time period are calculated. This method aims to capture the periodic characteristics of the occurrence of communication events to adapt to the changes in the frequency of communication events in different time periods.

[0125] Specifically, the system can divide a day into several time periods according to the actual application scenario. For example, in a data center environment, time can be divided into working hours (such as 9 am to 6 pm) and non-working hours (such as 6 pm to 9 am the next day). In each time period, the system counts the number of communication events in the sliding window Z and the time period length T, and calculates the Poisson parameter of the time period :

[0126] ;

[0127] Among them, the Poisson parameter of this time period is It represents the average occurrence rate of communication events in a unit of time. By calculating the Poisson parameters of each time period, the system can establish a basic probability model for the occurrence of communication events.

[0128] On this basis, the probability mass function of Poisson distribution is used to calculate the probability of communication events occurring within a specific time in the future. for:

[0129] ;

[0130] Where m is the number of events that occurred in a given time period and e is the base of the natural logarithm, which is approximately 2.718.

[0131] Based on the calculated λ, the probability of communication events occurring in the next time period and the expected number of occurrences can be predicted. For the prediction of time intervals, the exponential distribution can be used, and its probability density function is for:

[0132] ;

[0133] By solving the cumulative distribution function, the time interval for the next communication event is predicted.

[0134] For the two-state Markov chain model:

[0135] The status of the communication event is simplified into two states: "occurred" (state 1) and "not occurred" (state 0). The two states include state 0: the communication event did not occur; state 1: the communication event occurred.

[0136] Based on the historical data in the sliding window, calculate the state transition probability matrix : ;

[0137] in, It represents the probability of transitioning from state 0 to state 0, that is, the probability that the communication event will not occur at the next moment if it does not occur;

[0138] , represents the probability of transitioning from state 0 to state 1; It represents the probability of transitioning from state 1 to state 1, that is, the probability that the communication event still occurs at the next moment when it occurs; , represents the probability of transitioning from state 1 to state 0; state transition probability The calculation formula is:

[0139] ;

[0140] According to the current state and state transition probability, predict the probability of the communication event occurring at the next moment :

[0141] If the current state is state 0:

[0142] ;

[0143] If the current state is state 1:

[0144] ;

[0145] By combining the Poisson model and the two-state Markov chain model, the system can more accurately predict the probability of communication events in different time periods. The Poisson model provides the overall probability distribution of communication events, while the Markov chain model takes into account the continuity and dependency of communication events.

[0146] In practical applications, the system first uses the sliding window method to update the Poisson parameters in real time within each preset time period. And the state transition probability matrix P. Then, based on the current environmental data and communication event status, the probability of the communication event occurring at the next moment is predicted.

[0147] For example, suppose the Poisson parameter for the current time period is =0.5, indicating that a communication event occurs once every 2 units of time on average. If the current state is state 0, and the calculated , then the probability of a communication event occurring at the next moment is 30%. The system can decide whether to adjust the energy collection path based on this probability.

[0148] Based on the above prediction results, the system performs the following operations:

[0149] When the predicted probability is high: If When the set threshold (e.g., 50%) is exceeded, the system considers that a communication event may be about to occur. At this time, the conversion and storage unit 12 activates the main energy collection path according to the energy collection path adjustment instruction, adjusts the input parameters of the photovoltaic component 111, the thermoelectric component 112, and the radio frequency receiving component 113, and improves the energy collection efficiency. For example, the operating current of the photovoltaic component 111 is increased in advance, the load of the thermoelectric component 112 is optimized, or the operating mode of the radio frequency receiving component 113 is switched.

[0150] When the predicted probability is low: If If the energy level is lower than the set threshold, the system maintains the current energy collection mode, the conversion and storage unit 12 remains in a low-power sleep state, and only performs basic environmental data monitoring.

[0151] In this way, the system can dynamically adjust the energy collection strategy, improve energy utilization efficiency, and ensure that there is sufficient energy to supply the scattering communication unit 13 when a communication event occurs.

[0152] In addition, in order to adapt to changes in the environment and communication patterns, the system will regularly update the Poisson parameters and state transition probability matrix. For example, at the end of each sliding window, the communication event data is re-counted and new parameter values ​​are calculated. This dynamic update mechanism makes the system's prediction model adaptive and can promptly reflect changes in the regularity of communication events.

[0153] During the implementation process, the microcontroller in the RFID tag 10 is responsible for data collection, statistics and calculation. Since the computing and storage resources of the RFID tag 10 are limited, the improved probability model has a small amount of calculation and is suitable for implementation at the tag end. In order to reduce power consumption, the microcontroller can enter a sleep mode after completing the calculation, waiting for the next data collection or communication event to occur.

[0154] In summary, the present invention divides time into multiple preset time periods, combines Poisson distribution and two-state Markov chain model, and constructs an improved probability model for predicting the probability of occurrence of communication events. The model has small calculation amount, is easy to implement, can effectively guide the adjustment of energy collection path, improve the energy utilization efficiency of the system, and meet the needs of low-power RFID group scanning system.

[0155] As an optional implementation manner, the generating the energy collection path adjustment instruction includes:

[0156] Calculate the activation function E(t), where E(t) reflects the energy state and communication requirements of the low-power RFID group scanning system at time t;

[0157] When the excitation function E(t) reaches the preset trigger condition, an energy collection path adjustment instruction is generated to adjust the input parameters of each energy collection component and switch the energy collection mode;

[0158] When the excitation function E(t) does not reach the trigger condition, the current energy collection mode is maintained, and the conversion and storage unit 12 remains in a low-power sleep mode.

[0159] In the present invention, in order to realize intelligent adjustment of the energy collection path, an adjustment method based on the excitation function E(t) is proposed. The method aims to comprehensively reflect the energy state and communication requirements of the system at time t, and by calculating the excitation function E(t), it is determined whether it is necessary to generate an energy collection path adjustment instruction, thereby optimizing the energy collection strategy, meeting the communication requirements and saving energy.

[0160] In the specific implementation, first, the system calculates the activation function E(t) in real time during operation. The activation function E(t) is a comprehensive indicator used to evaluate factors such as the energy reserve of the system at the current moment, the predicted probability of the occurrence of communication events, and the energy consumption rate.

[0161] As an optional implementation, the calculation formula of the excitation function E(t) is:

[0162] ;

[0163] in is the predicted probability of a communication event occurring at time t, reflecting the urgency of the communication demand. This value is calculated using the aforementioned probability model and sliding window method, taking into account historical communication event data and current environment data;

[0164] is the amount of energy stored at time t, which represents the current electrical energy in the energy storage element (such as a supercapacitor or microbattery);

[0165] is the maximum capacity of energy storage, used to normalize the current energy storage; is the energy consumption rate at time t, which represents the energy consumption of the system in the current mode, including the basic system maintenance power consumption and potential communication power consumption.

[0166] , , The weight parameters are set according to system requirements to balance the impact of communication requirements, energy reserves and energy consumption on the activation function. These parameters can be optimized through experiments or simulations to meet the needs of different application scenarios.

[0167] When the activation function E(t) reaches the preset trigger condition (for example, E(t) is greater than a certain threshold ), the system considers it necessary to adjust the energy collection path to meet the upcoming communication needs. At this time, the conversion and storage unit 12 generates an energy collection path adjustment instruction, specifically performing the following operations:

[0168] According to the current environmental data and energy demand, the working parameters of each energy collection component (photovoltaic component 111, thermoelectric component 112, and radio frequency receiving component 113) are dynamically adjusted. For example, the working current of the photovoltaic component 111 is increased to optimize its working point; the load matching of the thermoelectric component 112 is adjusted; and the resonant frequency of the radio frequency receiving component 113 is tuned to adapt to the strongest radio frequency signal band in the environment.

[0169] Switch the system from low-power sleep mode to high-efficiency energy collection mode. At this point, the energy management circuit enables more functional modules to improve energy collection and conversion efficiency. For example, enable the MPPT (maximum power point tracking) algorithm to track the optimal energy collection state in real time; increase the sampling frequency and control accuracy to ensure that the energy collection components always work in the best state.

[0170] While adjusting the energy collection path, the system pre-allocates energy to the scattering communication unit 13 according to the predicted communication demand, ensuring that the scattering communication unit 13 can obtain sufficient power when a communication event actually occurs, thereby ensuring the reliability and timeliness of communication.

[0171] When the excitation function E(t) does not reach the trigger condition, the system determines that the current energy state and communication requirements do not require adjustment of the energy collection path. At this time, the current energy collection mode is maintained, and the conversion and storage unit 12 remains in a low-power sleep state. The system only retains basic environmental data monitoring and necessary energy collection functions to maintain the basic operation of the system with the lowest energy consumption.

[0172] It should be noted that the calculation of the activation function E(t) and the setting of the triggering conditions are crucial to the optimization of system performance. In order to ensure the accuracy of the activation function, the weight parameters need to be adjusted according to the actual operation conditions and historical data. , , Make dynamic adjustments.

[0173] In the specific implementation, firstly, the weight parameters are determined based on the actual application scenarios and performance requirements of the system. , , These parameters reflect the relative importance of the probability of communication events, energy storage, and energy consumption rate in the incentive function. For example, in scenarios with high communication reliability requirements, In the case of relatively tight energy supply, it can be appropriately increased or reduce The value of .

[0174] Then, based on historical data and experimental results, statistical analysis is performed on the E(t) values ​​of the system in different states. Through simulation or actual testing, the values ​​of the system's activation function under various environmental conditions, energy states, and communication requirements are collected. For example, the typical value of E(t) is recorded when a communication event is about to occur and energy reserves are insufficient; and the range of E(t) values ​​is recorded when energy is sufficient but communication requirements are low.

[0175] Next, statistical methods such as probability distribution analysis and threshold judgment theory are used to determine The best value of .

[0176] Exemplarily, the collected E(t) data is plotted into a probability distribution curve to observe its distribution characteristics under different states. Typically, the value of E(t) will present a bimodal or multimodal distribution, corresponding to different system states.

[0177] Select a threshold that can effectively distinguish between states where the system needs to adjust the energy harvesting path and states where it does not need to adjust. This threshold should be located at the intersection of the E(t) value distributions of different states to maximize the accuracy of the discrimination.

[0178] The selected threshold can be evaluated by cross-validation or hold-out method The threshold is adjusted to achieve the best system performance, ensuring that the energy harvesting path is adjusted in time when needed and avoiding unnecessary energy consumption when not needed.

[0179] In addition, in some application scenarios, It can be adjusted dynamically according to environmental changes and system operation status. For example, when the system detects that the environmental energy supply is sufficient, it can appropriately increase to reduce the energy consumption caused by frequent adjustments; conversely, when the energy supply is insufficient, the threshold is lowered to improve the system's responsiveness to communication needs.

[0180] Thus, the threshold is set to an optimal value that can balance the system energy efficiency and communication requirements. In this way, when the activation function E(t) exceeds The system will generate energy collection path adjustment instructions only when it is really needed, ensuring that energy adjustment is performed only when it is really needed, avoiding unnecessary energy waste.

[0181] In practical applications, such as in the management of optical fiber networks in data centers or communication base stations, the RFID tag 10 can achieve efficient use of energy through the above mechanism. When it is predicted that an optical fiber connection detection or configuration change is about to occur, the RFID tag 10 actively adjusts the energy collection path to ensure that there is enough energy for communication when the operation occurs. After the operation is completed, the system automatically returns to low power consumption mode, extending the working time of the equipment and reducing maintenance costs.

[0182] In this way, the present invention realizes intelligent adjustment of the energy collection path by calculating the excitation function E(t), effectively balances the energy state and communication requirements of the system, and improves the performance and reliability of the low-power RFID group scanning system.

[0183] See also Figure 3 , Figure 3 A flowchart of a multipath signal processing method provided for the implementation of the present invention, as an optional implementation, the multipath signal processing method includes S101 to S103, wherein:

[0184] S101: Under-sampling the received backscattered signal to obtain compressed sampling data of the signal;

[0185] S102: Reconstructing the original signal based on the compressed sampling data using a sparse reconstruction algorithm;

[0186] S103: demodulating and analyzing the reconstructed original signal to extract data information of each RFID tag 10;

[0187] Wherein, the sparse reconstruction algorithm includes:

[0188] Based on a preset sparse basis function, the compressed sampling data is mapped to a sparse domain to establish a sparse representation model of the signal; wherein the sparse basis function is a discrete cosine transform basis;

[0189] Based on the iterative hard threshold method, the sparse representation model is solved to reconstruct the original signal. The iterative hard threshold method performs the following steps in each iteration:

[0190] Calculate the gradient of the current iteration and obtain the update direction of the signal;

[0191] Perform a hard threshold operation on the gradient update result, retaining only the largest number of coefficients specified by the sparsity parameter, and setting the rest of the coefficients to zero;

[0192] Using the coefficients processed by hard threshold, an updated signal estimate is obtained through inverse transformation, and this is used as a reference for the reconstructed signal in the next iteration;

[0193] Repeat the above iterative process until the preset convergence condition is met or the maximum number of iterations is reached.

[0194] In the present invention, a multipath signal processing method based on undersampling and sparse reconstruction is disclosed for the backscattered signal received by the RFID reader 20 to improve the signal processing efficiency and accuracy, especially in a multi-tag environment, it can effectively distinguish and extract the data information of each RFID tag 10.

[0195] In a specific implementation, an analog-to-digital converter (ADC) with an adjustable sampling rate and a corresponding RF receiving channel are configured in the front-end signal receiving path of the RFID reader 20. Since the system adopts energy collection and low power consumption strategies, after the RFID reader 20 actively transmits the RF excitation signal, the antenna 131 receives the backscattered signals from multiple RFID tags 10. These tag signals are often superimposed together and have complex waveforms and propagation paths due to the multipath effect.

[0196] In this embodiment, the intelligent signal processing unit 23 of the RFID reader 20 performs undersampling on the superimposed signal at the receiving end, that is, extracts compressed sampled data at a sampling rate of one-half or lower of the conventional Nyquist sampling rate. The selection of the undersampling rate can be optimized according to the actual number of tags, the tag data rate, and the hardware resource constraints, for example, using 0.5 to 0.7 times the Nyquist rate for testing and optimization.

[0197] When the RFID reader 20 receives modulated backscatter signals from multiple RFID tags 10, it samples the signals at a frequency lower than the Nyquist frequency by controlling the sampling clock of the ADC. After a preset sampling window (e.g., a sampling time of several milliseconds to tens of milliseconds, depending on the tag communication protocol and rate setting), a set of under-sampled data sequences is obtained.

[0198] This data sequence is denoted as y(t) with a length of M (M << N), where N is the number of samples required for sampling at the Nyquist rate. In this step, the RFID reader 20 does not immediately perform conventional demodulation on the signal. Instead, the obtained samples are first input as the observed data in the compressive sensing framework into the subsequent sparse reconstruction module.

[0199] A set of sparse basis functions, such as the discrete cosine transform (DCT) basis, is pre-stored in the intelligent signal processing unit 23.

[0200] Furthermore, an assumption is made about the potential form of the original signal: The RFID backscatter signal after multipath superposition is sparse in a certain transform domain (such as the DCT domain), that is, only a few coefficients are non-zero or significant values in this domain.

[0201] The undersampled data y is expressed as:

[0202] ;

[0203] where is the undersampling matrix constructed according to the sampling strategy, is the sparse coefficient vector obtained under the DCT basis. On this basis, the original signal satisfies , is the inverse transform matrix of the DCT basis. That is, the original signal can be reconstructed in the time domain by performing an inverse transform on the sparse coefficient vector .

[0204] At this time, the problem is transformed into solving , given y and to reconstruct the original time-domain signal .

[0205] In a specific implementation, as the undersampling matrix is usually directly determined by the sampling strategy of the reader receiving channel.

[0206] For example, under the random sparse sampling strategy, the sampling control logic of the reader randomly selects M sampling points within a signal duration (corresponding to N samples for Nyquist sampling). The selection of these points can be generated by a random number generator or a pseudo-random sequence. Exemplarily, if N = 1000 and M = 300, then among the originally 1000 time points, the algorithm randomly selects 300 points for sampling. In the corresponding matrix (with a size of ), for the column positions of the selected 300 sample points, the corresponding rows are 1, and the remaining columns are 0. For example, if the points with indices 5, 27, 100... are selected, then The corresponding rows of are 1 in the 5th, 27th, 100th, etc., and 0 in the remaining columns. In this way, The specific locations where observations are extracted from the complete signal are determined.

[0207] For another example, for the sake of hardware implementation simplification or specific protocol requirements, a fixed interval sample selection strategy can be adopted. A sample can be taken every several points. Still taking N=1000 as an example, to obtain M=200 sampling points, one point can be selected every 5 points, so that a total of 1000 / 5=200 sampling points are finally selected. At this time, For one The matrix has a value of 1 in the specific index column corresponding to each row (for example, the first row corresponds to the first time point, the second row corresponds to the sixth time point, the third row corresponds to the eleventh time point, and so on), and a value of 0 in the unselected columns. This regular construction makes Has a simple periodic structure.

[0208] For example, when the reader has a certain prior knowledge of the channel characteristics, or has an estimate of the signal structure in the early stage, an adaptive strategy can be adopted. At the beginning of the system startup, the reader collects a small segment of the complete signal at a higher sampling rate, and analyzes the signal distribution, energy concentration segment, and tag data characteristics. Based on the analysis results, the time segment with high information content or the key frequency interval is determined. In the formal working stage, these segments are mainly sampled, while the segments with low information content are sampled in small amounts or not at all.

[0209] For example, if the initial analysis finds that the main energy of the original signal is concentrated at certain moments, then When the columns corresponding to these moments are selected as 1, the sampling points in the useless information period are reduced. The construction of not only considers the sparsity, but is also related to the actual channel characteristics and signal features to obtain higher quality reconstruction results at a lower sampling rate.

[0210] In summary, whether it is random, regular interval or adaptive strategy, The composition of can be regarded as selecting M actual sampling positions from N potential sampling positions according to the established strategy, and marking them as 1 in the corresponding columns, thus clarifying the method of obtaining under-sampling data. Finally, when the reader performs data collection, The relationship between the undersampled data y and the original signal s can be described in matrix form, thus providing a basis for subsequent sparse reconstruction.

[0211] In the specific implementation, the sparse reconstruction algorithm uses the iterative hard thresholding (IHT) method. The steps of the IHT algorithm are as follows:

[0212] (1) Initialization:

[0213] The initial value of the reconstruction coefficient vector Set to zero vector and set the maximum number of iterations (e.g. 50-100 times) and convergence criterion ε (e.g. when the difference between adjacent iterations is lower than a certain threshold, it is considered converged). At the same time, an upper limit value is set for the sparsity parameter S according to the actual system situation (e.g. S is a number of sparsity thresholds related to the expected number of labels and the time domain length of the signal).

[0214] (2) Iteration steps:

[0215] In the kth iteration, execute:

[0216] a. Calculate the gradient update direction:

[0217] Based on the current estimated signal coefficient , calculate the residual On this basis, the gradient update is calculated as:

[0218] ;

[0219] Among them is The step size parameter can be tuned based on experience or experiment (e.g. μ is in the range of 0.5 to 1). express The transpose of .

[0220] b. Hard threshold operation: update the result Perform hard thresholding:

[0221] i. Sort the coefficients in x' by absolute value from large to small;

[0222] ii. Only the first S largest amplitude coefficients (i.e., the sparsity parameter S defined above) are retained, and the remaining coefficients are cleared to zero to obtain the thresholded .

[0223] c. Convergence determination:

[0224] Compare and If the difference is less than the preset threshold ε, or k has reached , the iteration is terminated; otherwise, continue to the next iteration.

[0225] Through multiple iterations, the When the iteration is completed, The sparse representation of the original signal in the DCT domain is retained. The sparse coefficient vector finally obtained after the iterative hard threshold method completes all preset iterative steps or reaches the convergence condition, that is, the final coefficient estimation result;

[0226] Furthermore, after completing the IHT iteration, the reconstructed original signal is obtained through inverse transformation:

[0227] ;

[0228] Obtained That is the backscattered signal approximately restored in the time domain.

[0229] Subsequently, the intelligent signal processing unit 23 demodulates and analyzes the reconstructed time domain signal. According to the modulation method adopted by the RFID tag 10 (for example, a specific coding scheme such as FM0 or Miller coding), the system performs bandpass filtering, coherent demodulation, bit synchronization and decoding on the reconstructed signal. By distinguishing the modulation characteristics and the specific code sequence, the independent data sequence of each RFID tag 10 is extracted from the reconstructed signal.

[0230] After compressed sensing and sparse reconstruction, the reconstructed signal waveform is properly synchronized and the decision threshold is adjusted to identify the unique ID of each tag and its additional information. The information of multiple tags is represented as a superposition of different delay and phase characteristics at the signal level. After sparse reconstruction, the data stream of each tag can be accurately distinguished by matching the specific time slot or specific code element position.

[0231] When multiple tag signals are superimposed in a multipath environment, the above-mentioned undersampling and sparse reconstruction method can effectively restore the main component signal under low power consumption and low sampling rate conditions, and significantly reduce the impact of noise and interference. After completing the analysis, the intelligent signal processing unit 23 structures the information of each tag and transmits it to the data management unit 30 for recording and logical association establishment.

[0232] In this way, the multi-path signal processing method described in this embodiment has the characteristics of strong operability, clear steps and adjustable parameters. In the actual system, the ADC sampling rate, iterative algorithm parameters and sparsity selection can be optimized to achieve stable and low-power multi-tag recognition and data reading and writing.

[0233] Through the above multipath signal processing method, the RFID reader 20 can accurately reconstruct and analyze the backscattered signal in a complex multipath environment and multi-tag situation, and extract the data information of each RFID tag 10. This improves the recognition rate and reliability of the system and meets the needs of efficient and accurate reading of a large number of RFID tags 10 in practical applications.

[0234] In addition, this method also has advantages in reducing hardware cost and power consumption. Due to the use of undersampling and sparse reconstruction technology, a low-rate analog-to-digital converter (ADC) and a lower sampling frequency can be used, reducing the power consumption and hardware complexity of the system. This is of great significance for RFID readers 20 that need to run continuously for a long time.

[0235] In summary, the multipath signal processing method provided by the present invention obtains compressed sampling data through undersampling, reconstructs the original signal using a sparse reconstruction algorithm and an iterative hard threshold method, and demodulates and analyzes the reconstructed signal, thereby successfully solving the problem of reading RFID signals in complex environments and improving the performance and application value of low-power RFID group scanning systems.

[0236] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0237] It should be understood that determining B based on A does not mean determining B only based on A. B can also be determined based on A and / or other information.

[0238] Those skilled in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed in the present invention can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0239] In the description of this specification, the description with reference to the terms "one embodiment", "example", "specific example", etc. means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representation of the above terms does not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or more embodiments or examples in a suitable manner.

[0240] The preferred embodiments of the present invention disclosed above are only used to help explain the present application. The preferred embodiments do not describe all the details in detail, nor do they limit the present application to specific implementation methods. Obviously, many modifications and changes can be made according to the content of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the present application, so that those skilled in the art can understand and use the present application well. The present application is limited only by the claims and their full scope and equivalents.

Claims

1. Low-power RFID group scanning system based on energy harvesting and backscattering technology, characterized by: include: A plurality of RFID tags are attached to the optical fiber port and the pigtail respectively; wherein each of the RFID tags comprises: a collection unit, used to collect radio frequency energy from the environment and convert it into electrical energy; a conversion and storage unit, used to convert and store the electrical energy using a preset energy storage element based on a maximum power point tracking method, and to power the RFID tag; a scattering communication unit, comprising an antenna and a signal reflection unit, wherein the antenna is used to receive a radio frequency excitation signal from an RFID reader, and the signal reflection unit is used to reflect a modulated signal containing tag information back to the RFID reader by backscattering; The RFID reader / writer includes: a radio frequency transmitting unit, used to send a radio frequency excitation signal to the RFID tag; a radio frequency receiving unit, used to receive a backscattered signal from the RFID tag; an intelligent signal processing unit, connected to the radio frequency receiving unit, and based on a multi-path signal processing method, analyzing the received backscattered signal and distinguishing signals from multiple RFID tags; A data management unit is used to collect, store and manage the read RFID tag information and automatically establish a logical association between the fiber port label and the pigtail label; The multipath signal processing method comprises: Under-sampling the received backscattered signal to obtain compressed sampling data of the signal; Using sparse reconstruction algorithm, the original signal is reconstructed based on the compressed sampling data; Demodulate and analyze the reconstructed original signal to extract the data information of each RFID tag; Wherein, the sparse reconstruction algorithm includes: Based on a preset sparse basis function, the compressed sampling data is mapped to a sparse domain to establish a sparse representation model of the signal; wherein the sparse basis function is a discrete cosine transform basis; Based on the iterative hard threshold method, the sparse representation model is solved to reconstruct the original signal. The iterative hard threshold method performs the following steps in each iteration: Calculate the gradient of the current iteration and obtain the update direction of the signal; Perform a hard threshold operation on the gradient update result, retaining only the largest number of coefficients specified by the sparsity parameter, and setting the rest of the coefficients to zero; Using the coefficients processed by hard threshold, an updated signal estimate is obtained through inverse transformation, and this is used as a reference for the reconstructed signal in the next iteration; Repeat the above iterative process until the preset convergence condition is met or the maximum number of iterations is reached.

2. The low-power RFID group scanning system based on energy collection and backscattering technology according to claim 1 is characterized in that: The collection unit is also used to collect light energy and heat energy from the environment; The radio frequency energy is collected from the radio wave energy in the environment through the radio frequency receiving component and converted into electrical energy; Collect light energy through photovoltaic modules and convert it into electrical energy; Collect ambient temperature difference energy through thermoelectric components and convert it into electrical energy; The electric energy collected by the radio frequency receiving component, the photovoltaic component and the thermoelectric component is input into the conversion and storage unit to perform electric energy integration and output control.

3. The low-power RFID group scanning system based on energy collection and backscattering technology according to claim 2 is characterized in that: The maximum power point tracking method comprises: Collect environmental data, combine it with historical communication event data, use the sliding window method to calculate the probability and time interval of communication events, and generate prediction results; Based on the prediction result, determining the triggering time and energy requirement of the communication event, and generating an energy collection path adjustment instruction; The energy collection path adjustment instruction includes adjusting input parameters of the energy collection component and triggering conditions for mode switching; the energy collection component includes a radio frequency receiving component, a photovoltaic component and a thermoelectric component; In response to the energy harvesting path adjustment instruction not triggering the switching condition, the conversion and storage unit enters a low-power sleep mode and only maintains low-frequency monitoring of environmental data; In response to the energy collection path adjustment instruction triggering a switching condition and a communication event about to occur, activating the main energy collection path, and adjusting input parameters of the radio frequency receiving component, the photovoltaic component, and the thermoelectric component; In response to the actual triggering of the communication event, the conversion and storage unit supplies power to the scattering communication unit at a first priority according to the allocation scheme determined by the energy collection path adjustment instruction; In response to the completion of the communication event, the default parameter configuration is restored based on the energy collection path adjustment instruction, the low power sleep mode is switched back, and the environment monitoring state is re-entered.

4. The low-power RFID group scanning system based on energy collection and backscattering technology according to claim 3 is characterized in that: The sliding window method includes: Set sliding windows to update and manage collected environmental data and historical communication event data in real time; In each sliding window, the number of communication events and the time interval are counted, and the average occurrence rate of communication events is calculated; An improved probability model is used to calculate the occurrence probability and predicted time interval of a communication event according to statistical results, wherein the probability model is a piecewise Poisson process or a two-state Markov chain model; Based on the probability of communication event occurrence and the predicted time interval, a prediction result for energy harvesting path adjustment is generated.

5. The low-power RFID group scanning system based on energy collection and backscattering technology according to claim 4 is characterized in that: The improved probability model includes: Divide the time into a plurality of preset time periods, consider the occurrence rate of communication events in each time period to be constant, and calculate the Poisson parameter of each time period; For the two states of communication event occurrence and non-occurrence, the probability of communication event occurrence at the next moment is predicted based on the state and state transition probability at the previous moment.

6. The low-power RFID group scanning system based on energy collection and backscattering technology according to claim 5 is characterized in that: The length of the sliding window and the sliding step size are adaptively adjusted according to the frequency of communication events and the rate of environmental change.

7. The low-power RFID group scanning system based on energy collection and backscattering technology according to claim 6 is characterized in that: The generating of the energy collection path adjustment instruction comprises: Calculate the activation function E(t), where E(t) reflects the energy state and communication requirements of the low-power RFID group scanning system at time t; When the excitation function E(t) reaches the preset trigger condition, an energy collection path adjustment instruction is generated to adjust the input parameters of each energy collection component and switch the energy collection mode; When the excitation function E(t) does not reach the trigger condition, the current energy collection mode is maintained, and the conversion and storage unit remains in a low-power sleep mode.

8. The low-power RFID group scanning system based on energy collection and backscattering technology according to claim 7 is characterized in that: The calculation formula of the activation function E(t) is: ； in is the predicted probability of the communication event occurring at time t, is the energy storage at time t, is the maximum capacity of energy storage, is the energy consumption rate at time t, The weight parameters are set according to the requirements of the low-power RFID group scanning system.

Images