Full-life traceability type digital shelf system of Internet of Things

Abstract

The invention provides a full-life traceability type digital shelf system based on the Internet of Things, and the system comprises a multi-source heterogeneous sensor group which is disposed at a raw material supply device, a production device, a logistics carrier, and an intelligent shelf, and the multi-source heterogeneous sensor group comprises: the production device is provided with a vibration spectrum sensor, and collects device vibration waveform data bound with a raw material batch number in real time; the logistics carrier is provided with a position sensor, and transport track GPS coordinates and corresponding time-space stamps are synchronously recorded; the intelligent goods shelf is provided with a pressure sensing array, and the commodity distribution form is monitored with 1 cm < 2 > grid precision. And a block chain causal evidence storage module. According to the full-life traceability type digital shelf system of the Internet of Things, the equipment vibration waveform is accurately collected and bound with the batch number of the raw material, so that the quality problem can be quickly traced to the raw material, the production quality management and control level is improved, high-precision space-time positioning and time synchronization are realized in logistics transportation, the transportation track can be accurately recorded, and the transportation efficiency is improved. Abnormal events are comprehensively traced, and logistics information transparency is guaranteed.

Description

Technical Field

[0001] The present invention relates to the technical field of digital shelf systems, and particularly to an Internet of Things full-life-cycle traceable digital shelf system. Background Art

[0002] In today's digital and intelligent business environment, the efficiency and transparency of commodity supply chain management are becoming increasingly crucial. The rise of Internet of Things technology provides new opportunities for optimizing all links of the supply chain. However, there are still many limitations in the current supply chain traceability system;

[0003] In contrast, the patent document with publication number CN112926703B discloses a method, device and system for processing commodity object display information. The system includes: a plurality of near-field communication tags at a plurality of preset positions on the shelf for storing first information, where the first information includes information associated with the near-field communication tag; a commodity object tag associated with a commodity object, which is used to be arranged at a position corresponding to the associated commodity object on the shelf, and the commodity object tag is correspondingly provided with a near-field communication card reader for reading the first information stored in the near-field communication tag through the card reader; a server for determining the position information of the commodity object associated with the commodity object tag on the shelf according to the received first information and the pre-stored second information; the second information includes position information related to the near-field communication tag. Through the embodiments of the present application, it is possible to automatically obtain the actual display situation of commodity objects on the shelves of physical stores.

[0004] Based on the above patent document and the prior art, traditional means for monitoring production equipment can often only obtain limited equipment status information, making it difficult to accurately and real-time collect subtle but important features of equipment vibration, and lacking an effective binding with raw material batch numbers. As a result, when quality problems occur, it is difficult to quickly and accurately trace back to the raw material batches and the corresponding production process information. In the warehousing and sales links, most existing shelf systems can only simply count the quantity of commodities, and cannot monitor the distribution form of commodities with high precision, resulting in great difficulties in real-time management and accurate calculation of inventory. At the same time, there is a lack of a hierarchical data display and traceability mechanism for consumers and regulators. Consumers are difficult to obtain comprehensive and intuitive commodity information, and regulators are also unable to efficiently conduct quality supervision and problem tracing.

[0005] In summary, there are many deficiencies in the prior art in terms of full-life-cycle traceability of commodities. There is an urgent need for an Internet of Things full-life-cycle traceable digital shelf system that can integrate the whole-link data of production, logistics, warehousing and sales, realize accurate traceability, reliable evidence storage and intelligent prediction. Summary of the Invention

[0006] The technical problem to be solved by the present invention is to overcome the defects existing in the prior art, and the present invention proposes a full-life traceable digital shelf system for the Internet of Things.

[0007] To solve the above technical problems, the technical solution adopted by the present invention is: The full-life traceable digital shelf system for the Internet of Things, including:

[0008] A multi-source heterogeneous sensor group, deployed in raw material supply, production equipment, logistics carriers and intelligent shelves. The multi-source heterogeneous sensor group includes: vibration spectrum sensors are set on production equipment, and acquisition terminals with image recognition and NFC readers are set in raw material supply to collect equipment vibration waveform data bound to raw material batch numbers; position sensors are set on logistics carriers to synchronously record the GPS coordinates of transportation trajectories and corresponding time stamps; a pressure sensing array is set on the intelligent shelf to monitor the commodity distribution form with a grid accuracy of 1 cm 2 grid accuracy;

[0009] The blockchain causal evidence storage module includes:

[0010] An intelligent contract trigger unit, when the vibration waveform data exceeds the equipment health threshold, automatically initiates a cross-chain verification request to a third-party testing agency;

[0011] A spatio-temporal anchor ledger, which aligns production vibration data, logistics trajectory coordinates and shelf pressure data according to the IEEE-1588v2 protocol to generate a joint hash value, and stores it as an event chain unit with a time stamp;

[0012] The trusted intelligent central module includes:

[0013] A virtual inventory backtracking unit, which constructs a physical constraint equation based on the gradient change of the shelf pressure distribution and outputs a verifiable inventory quantity with zero-knowledge proof;

[0014] A causal enhanced prediction unit, which uses the LSTM-TCN model for demand prediction. When the prediction error exceeds 5%, it reversely activates the holographic traceability module to obtain fault batch data, and dynamically updates the model weight parameters through gradient backpropagation;

[0015] The holographic traceability execution engine module includes:

[0016] Edge computing nodes, deployed at the intelligent shelf terminal, to perform biometric binding and scanning operations;

[0017] A hierarchical data reconstruction unit, which dynamically assembles the event chain according to the access permission:

[0018] Display the quality inspection report and the logistics trajectory heat map for consumer permissions;

[0019] The supervision authority opens the vibration spectrum, temperature and humidity fluctuation coupling analysis view, and calls the space-time anchor ledger to restore the equipment anomaly timeline.

[0020] Preferably, the vibration characteristic sensor adopts the resonance peak detection method, collects vibration waveforms with a frequency range of 10 Hz - 5 kHz, and binds them with the raw material batch number through the following formula: where B id is the batch identification code, A f is the characteristic frequency matrix, Ψ(t) is the time-domain vibration waveform function, is the integral operation of the time-domain vibration waveform function Ψ(t) in the time interval [t 1 , t 2 , and HASH is the hash function.

[0021] Preferably, the micro-pressure sensing array is deployed with a grid accuracy of 1 cm 2 , and adjacent sensor nodes use Kalman filtering for data fusion. Its pressure distribution function is:

[0022] where ω i is the weight coefficient of the i-th node, σ i is the standard deviation of the Gaussian distribution, P(x, y) is the pressure distribution function, which is used to describe the pressure distribution at the position of the plane coordinates (x, y). n is the number of sensor nodes, (x i , y i ) is the position of the i-th sensor node in the plane coordinates. The space-time positioning device synchronously records the GPS coordinates and the Beidou satellite timing signal, and the time synchronization error is less than 1 ms.

[0023] Preferably, the IEEE-1588v2 protocol is used to align the clocks of each node, and a space-time coordinate system conversion model is established where α is the clock drift coefficient, β is the reference time offset, γ k is the spatial position compensation parameter, T g is the global time, T l is the local time, that is, the clock time of each node in the system. m is the number of spatial position-related parameters.

[0024] Preferably, the virtual inventory back-inference unit establishes a physical constraint equation between the pressure gradient and the inventory quantity, and through the physical constraint relationship between the pressure gradient vector and the commodity density, an integral equation between the time derivative of the shelf height distribution function and the pressure gradient is established, where the pressure gradient direction is orthogonal to the normal vector of the shelf surface.

[0025] Preferably, a zero-knowledge proof protocol is used to generate a verifiable inventory. The verification process includes an equivalence test of the linear homomorphic commitment of the inventory secret parameter and the verification coefficient in the prime field.

[0026] Preferably, the causal enhancement prediction unit includes a hybrid structure of a long short-term memory network and a temporal convolutional layer. The temporal convolutional layer performs weighted summation calculation on the input data through the dilation coefficient and the number of convolutional kernels. When the mean absolute error between the predicted value and the actual value exceeds 5% of the maximum actual value, backward tracing is triggered, and the model weights are updated using a gradient descent algorithm with L1 constraint, where the constraint strength increases linearly with the number of faulty batches.

[0027] Preferably, the input of the long short-term memory network is multi-dimensional data including historical inventory, sales data, and time series features. The gating mechanism is used to selectively remember and forget information to capture long-term dependencies in the data.

[0028] Preferably, when generating the heat map of the logistics track for consumer permission display, the hierarchical data reconstruction unit uses the kernel density estimation method to adaptively adjust the bandwidth according to the distribution of the logistics track points to intuitively display the density of the logistics track.

[0029] Preferably, when the holographic traceability execution engine module receives a query request for supervision authority, during the process of calling the spatio-temporal anchor ledger to restore the device abnormal time axis, combined with the device vibration spectrum abnormality quantification index and the device health equation, the time point of the device abnormality occurrence and the remaining service life assessment are determined, and reasonable thresholds for key indicators are set. When the sensor data is abnormal, the system can immediately send out a warning signal to remind relevant personnel to handle it in time;

[0030] According to the data analysis results, based on the device status data and the product quality traceability information.

[0031] Compared with the prior art, the beneficial effects of the present invention include: through the full-life traceable digital shelf system of the Internet of Things, in the production link, the device vibration waveform is accurately collected and bound to the raw material batch number, so that quality problems can be quickly traced back to the raw materials, improving the production quality control level. In the logistics transportation, high-precision spatio-temporal positioning and time synchronization can accurately record the transportation track, comprehensively trace abnormal events, and ensure the transparency of logistics information. In warehousing and sales, the inventory is calculated by combining the physical constraint equation, improving the accuracy of inventory management, accurately predicting demand, dynamically updating the model weights, and enhancing the supply chain response speed, forming a complete and reliable commodity full-life cycle traceability system, and comprehensively optimizing the commodity supply chain management. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] The disclosure of the present invention will be described with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of the present invention. In the drawings, the same reference numerals are used to refer to the same components. Among them:

[0033] Figure 1 Schematically shows a flow diagram of a full-life traceable digital shelf system for the Internet of Things according to an embodiment of the present invention.

[0034] Figure 2 Schematically shows a logical decision diagram of a full-life traceable digital shelf system for the Internet of Things according to an embodiment of the present invention. Detailed implementation manners

[0035] It is easy to understand that according to the technical solution of the present invention, without changing the essence of the present invention, those of ordinary skill in the art can propose various replaceable structural ways and implementation ways. Therefore, the following detailed implementation manners and the accompanying drawings are only exemplary descriptions of the technical solution of the present invention, and should not be regarded as all of the present invention or as a limitation or restriction on the technical solution of the present invention.

[0036] According to an embodiment of the present invention in combination with Figure 1-2 Shown. A full-life traceable digital shelf system for the Internet of Things includes:

[0037] A multi-source heterogeneous sensor group, deployed in raw material supply, production equipment, logistics carriers and intelligent shelves. Use an image recognition sensor to scan the QR code or barcode on the raw material package to obtain the origin, supplier name and unique identification code of the raw material. The supplier pre-stores the raw material composition analysis report in the form of an electronic document and writes it into the NFC tag. The acquisition terminal obtains the composition data through the NFC reader, including the proportion of the main components, to ensure that the raw material composition information enters the traceability system. With the help of NFC technology, the acquisition terminal reads the quality inspection report of the raw material. An Internet of Things tracking device is set on the raw material transportation. The device real-time collects the position information, transportation time and temperature and humidity parameters in the carriage during the transportation process. At the same time, the position information is obtained through the Beidou positioning system. The transportation data will be regularly transmitted to the traceability system and integrated with other raw material information.

[0038] The multi-source heterogeneous sensor group includes: a vibration spectrum sensor is set on the production equipment, and an acquisition terminal with an image recognition and NFC reader is set on the raw material supply to collect the device vibration waveform data bound to the raw material batch number; a position sensor is set on the logistics carrier to synchronously record the GPS coordinates of the transportation track and the corresponding space-time stamp; a pressure sensing array is set on the intelligent shelf to monitor the commodity distribution form with a grid accuracy of 1 cm 2 Grid accuracy to monitor the commodity distribution form;

[0039] The vibration spectrum sensor captures the vibration signal of the device and converts it into waveform data using a built-in algorithm. The position sensor of the logistics vehicle uses a high-precision GPS module, combines satellite positioning technology, real-time tracks the position change of the vehicle, and synchronously records the timestamp to ensure the accuracy and temporal coherence of the transportation trajectory information. The pressure sensing array of the intelligent shelf consists of multiple micro pressure sensors, arranged in a 1 cm 2 grid pattern, which can finely sense the pressure changes at different positions on the shelf. In the production process, the vibration spectrum sensor is bound to the raw material batch number to collect data, which helps to quickly locate product quality problems caused by abnormal production equipment, trace back to the corresponding raw material batch, and improve the production quality control ability. The synchronized recording of the position and time of the logistics vehicle makes the logistics transportation process completely transparent, facilitating the timely discovery of problems such as transportation delays and route deviations, and ensuring the timely and accurate delivery of goods. The high-precision pressure sensing array of the intelligent shelf can real-time monitor the placement and quantity changes of goods, providing accurate data support for inventory management.

[0040] The vibration feature sensor adopts the resonance peak detection method, collects vibration waveforms with a frequency range of 10 Hz - 5 kHz, and is bound to the raw material batch number through the following formula: where B id is the batch identification code, A f is the characteristic frequency matrix, Ψ(t) is the time-domain vibration waveform function, is the integral operation of the time-domain vibration waveform function Ψ(t) over the time interval [t 1 , t 2 , HASH is the hash function, the micro pressure sensing array is deployed with a 1 cm 2 grid accuracy, and adjacent sensor nodes use Kalman filtering for data fusion. Its pressure distribution function is:

[0041] where ω i is the weight coefficient of the i-th node, σ i is the standard deviation of the Gaussian distribution, P(x, y) is the pressure distribution function, used to describe the pressure distribution at the position of the plane coordinates (x, y), n is the number of sensor nodes, (x i , y i ) is the position of the i-th sensor node in the plane coordinates, and the spatio-temporal positioning device synchronously records the GPS coordinates and Beidou satellite time signal, and the time synchronization error is less than 1 ms.

[0042] The blockchain causal evidence storage module includes: uploading raw material data to the chain, sorting out the raw material information, encrypting these data using an encryption algorithm to ensure the security and integrity of the data, and uploading the encrypted data to the spatio-temporal anchor ledger of the blockchain through a smart contract;

[0043] Data association and traceability. When generating the combined hash value, raw material data, production vibration data, logistics track coordinates, and shelf pressure data are included in the calculation scope. When full-life traceability is required, by accessing the spatio-temporal anchor ledger, all-process data of raw materials from supply to production, logistics, sales, etc. can be quickly queried based on the combined hash value, truly realizing one-stop full-life query.

[0044] Intelligent contract trigger unit. When the vibration waveform data exceeds the device health threshold, it automatically initiates a cross-chain verification request to a third-party inspection agency. The intelligent contract has pre-set the device health threshold, which is determined based on the historical operation data of the device. Once the vibration waveform data exceeds the set threshold, the intelligent contract automatically sends a cross-chain verification request to the certified third-party inspection agency through the communication protocol of the blockchain network. The request contains detailed device vibration data and relevant information, automatically triggering cross-chain verification, enabling timely professional detection and evaluation of the abnormal state of the device, ensuring accurate diagnosis of device problems, avoiding production interruptions or product quality problems caused by device failures, and at the same time enhancing the credibility and authority of the data by virtue of the public credibility of the third-party inspection agency.

[0045] Spatio-temporal anchor ledger. Align the production vibration data, logistics track coordinates, and shelf pressure data according to the IEEE-1588v2 protocol to generate a combined hash value, and store it as an event chain unit with spatio-temporal stamps;

[0046] Adopt the IEEE-1588v2 protocol to align the clocks of each node and establish a spatio-temporal coordinate system conversion model where α is the clock drift coefficient, β is the reference time offset, and γ k is the spatial position compensation parameter, T g is the global time, and T l is the local time, that is, the clock time of each node in the system itself. m is the number of spatial position-related parameters.

[0047] Trusted intelligent central module, including:

[0048] The virtual inventory inversion unit constructs a physical constraint equation based on the gradient change of the shelf pressure distribution and outputs a verifiable inventory quantity with zero-knowledge proof. The virtual inventory inversion unit establishes a physical constraint equation between the pressure gradient and the inventory quantity. Through the physical constraint relationship between the pressure gradient vector and the commodity density, an integral equation of the time derivative of the shelf height distribution function and the pressure gradient is established. The pressure gradient direction is orthogonal to the normal vector of the shelf surface. The zero-knowledge proof protocol is used to generate a verifiable inventory quantity. The verification process includes the equivalence test of the linear homomorphic commitment of the inventory secret parameter and the verification coefficient in the prime number field. When constructing the physical constraint equation, the physical constraint equation can accurately calculate the inventory based on the shelf pressure, improving the efficiency and accuracy of inventory management. The zero-knowledge proof protocol protects business secrets while ensuring the credibility of inventory data and meeting the verification requirements of inventory data in different scenarios.

[0049] The causal enhanced prediction unit uses the LSTM-TCN model for demand prediction. When the prediction error exceeds 5%, it reversely activates the holographic traceability module to obtain the faulty batch data and dynamically updates the model weight parameters through gradient backpropagation. The causal enhanced prediction unit includes a hybrid structure of a long short-term memory network and a temporal convolutional layer. The temporal convolutional layer performs weighted summation calculations on the input data through the dilation coefficient and the number of convolutional kernels. When the mean absolute error between the predicted value and the actual value exceeds 5% of the maximum actual value, reverse tracing is triggered. The model weight update uses the gradient descent algorithm with L1 constraint, where the constraint strength increases linearly with the number of faulty batches. The input of the long short-term memory network is multi-dimensional data including historical inventory quantity, sales data, and time series features. The gating mechanism is used to selectively remember and forget information to capture long-term dependencies in the data. The LSTM-TCN model combines the ability of the long short-term memory network to capture long-term dependencies and the efficient processing ability of the temporal convolutional network (TCN) for time series data. During the model training process, a large amount of historical data is used for iterative optimization to continuously adjust the model weight parameters. When the prediction error exceeds the set threshold, the reverse tracing mechanism obtains detailed faulty batch data through interaction with the holographic traceability module, including equipment parameters and raw material information in the production process. Accurate demand prediction helps enterprises reasonably arrange production and inventory, reducing operating costs. The reverse tracing and dynamic model update mechanism can adjust the model in a timely manner according to the actual situation, improving the accuracy of prediction and providing strong support for the root cause analysis of problems.

[0050] The holographic traceability execution engine module includes:

[0051] Edge computing nodes are deployed at the terminals of intelligent shelves to perform biometric binding and scanning operations. The edge computing nodes can quickly and accurately identify biometric information and bind it to the identification code of the commodity. The scanning operation uses two-dimensional code or RFID identification technology to ensure the rapid reading and accurate transmission of commodity information;

[0052] The hierarchical data reconstruction unit dynamically assembles the event chain according to the access authority. When generating the logistics track heat map for consumer authority display, the hierarchical data reconstruction unit uses the kernel density estimation method to adaptively adjust the bandwidth according to the distribution of logistics track points, so as to intuitively display the density of the logistics track.

[0053] The quality inspection report of consumer authority display and the logistics track heat map. When generating the logistics track heat map, the kernel density estimation method analyzes the spatial distribution of logistics track points and automatically adjusts the bandwidth parameter to adapt to the change of track density in different regions. For the data display of supervision authority, combined with professional data analysis tools and visualization interfaces, complex vibration spectrum and temperature and humidity fluctuation data are presented in the form of intuitive charts. The device abnormality quantification index is obtained by comprehensively calculating multiple characteristic parameters of the vibration spectrum.

[0054] The supervision authority opens the coupling analysis view of vibration spectrum and temperature and humidity fluctuation, and calls the space-time anchor ledger to restore the device abnormality time axis. When the holographic traceability execution engine module receives the query request of the supervision authority, in the process of calling the space-time anchor ledger to restore the device abnormality time axis, combined with the device vibration spectrum abnormality quantification index and the device health equation, determine the time point of device abnormality occurrence and the remaining service life assessment, set reasonable thresholds for key indicators, and when the sensor data is abnormal, the system can immediately send out a warning signal to remind relevant personnel to deal with it in time;

[0055] According to the data analysis results, based on the device status data and product quality traceability information.

[0056] The technical scope of the present invention is not limited to the content in the above description. Those skilled in the art can make various deformations and modifications to the above embodiments without departing from the technical idea of the present invention, and these deformations and modifications should all fall within the protection scope of the present invention.

Claims

1. The full life cycle traceability digital shelf system of the Internet of Things is characterized by: include: The multi-source heterogeneous sensor group is deployed in raw material supply, production equipment, logistics vehicles and smart shelves. The multi-source heterogeneous sensor group includes: the production equipment is equipped with a vibration spectrum sensor; the raw material supply is equipped with a collection terminal with image recognition and NFC reader to collect equipment vibration waveform data bound to the raw material batch number; the logistics vehicle is equipped with a position sensor to synchronously record the transportation trajectory GPS coordinates and corresponding time and space stamps; the smart shelf is equipped with a pressure sensor array to measure the pressure of the smart shelf with a 1cm 2 Grid precision monitors commodity distribution patterns; The blockchain causal evidence storage module includes: The smart contract trigger unit automatically initiates a cross-chain verification request to a third-party testing agency when the vibration waveform data exceeds the device health threshold; The spatiotemporal anchor ledger aligns the production vibration data, logistics trajectory coordinates, and shelf pressure data according to the IEEE-1588v2 protocol to generate a joint hash value, which is stored as an event chain unit with a time and space stamp; Trusted intelligent hub module, including: The virtual inventory reverse calculation unit builds physical constraint equations based on the gradient change of shelf pressure distribution and outputs verifiable inventory with zero-knowledge proof; The causal enhancement prediction unit uses the LSTM-TCN model to predict demand. When the prediction error exceeds 5%, the holographic tracing module is activated in reverse to obtain the fault batch data, and the model weight parameters are dynamically updated through gradient back propagation. Holographic traceability execution engine module, including: Edge computing nodes are deployed at smart shelf terminals to perform biometric binding and code scanning operations; Hierarchical data reconstruction unit, dynamically assembling event chains based on access rights: Consumers can display quality inspection reports and logistics trajectory heat maps; The regulatory authority opens up the vibration spectrum and temperature and humidity fluctuation coupling analysis views, and calls the space-time anchor point ledger to restore the equipment abnormality timeline.

2. The full life cycle traceability digital shelf system of the Internet of Things according to claim 1 is characterized in that: The vibration characteristic sensor adopts the resonance peak detection method to collect the vibration waveform with a frequency range of 10Hz-5kHz, and binds it with the raw material batch number through the following formula: Among them, B id is the batch identification code, A f is the characteristic frequency matrix, Ψ(t) is the time domain vibration waveform function, To perform an integration operation on the time domain vibration waveform function Ψ(t) in the time interval [t1, t2], HASH is a hash function.

3. The full life cycle traceability digital shelf system of the Internet of Things according to claim 2 is characterized in that: The micro pressure sensing array is 1cm 2 Grid precision deployment, adjacent sensor nodes use Kalman filtering for data fusion, and its pressure distribution function is: Among them, ω i is the weight coefficient of the i-th node, σ i is the standard deviation of Gaussian distribution, P(x,y) is the pressure distribution function, which is used to describe the pressure distribution at the plane coordinate (x,y), n is the number of sensor nodes, (x i ,y i ) is the position of the i-th sensor node in the plane coordinates. The space-time positioning device synchronously records the GPS coordinates and the Beidou satellite timing signal, and the time synchronization error is less than 1ms.

4. The full life cycle traceability digital shelf system of the Internet of Things according to claim 1 is characterized in that: Use IEEE-1588v2 protocol to align the clocks of each node and establish a space-time coordinate system conversion model Where α is the constant drift coefficient, β is the reference time offset, and γ k is the spatial position compensation parameter, T g is the global time, T l is the local time, i.e. the clock time of each node in the system. m is the number of spatial position related parameters.

5. The full life cycle traceability digital shelf system of the Internet of Things according to claim 1 is characterized in that: The virtual inventory reverse calculation unit establishes a physical constraint equation between pressure gradient and inventory quantity, and establishes an integral equation between the time derivative of the shelf height distribution function and the pressure gradient through the physical constraint relationship between the pressure gradient vector and the commodity density, wherein the direction of the pressure gradient is orthogonal to the normal vector of the shelf surface.

6. The full life cycle traceability digital shelf system of the Internet of Things according to claim 5 is characterized in that: A zero-knowledge proof protocol is used to generate verifiable inventory. The verification process includes checking the equivalence of the linear homomorphic commitments between the secret parameters of the inventory and the verification coefficients on the prime field.

7. The full life cycle traceability digital shelf system of the Internet of Things according to claim 1 is characterized in that: The causal enhanced prediction unit includes a hybrid structure of a long short-term memory network and a temporal convolutional layer, wherein the temporal convolutional layer performs weighted summation calculation on the input data through the expansion coefficient and the number of convolution kernels, and triggers reverse tracing when the mean absolute error between the predicted value and the actual value exceeds 5% of the maximum actual value. The model weight is updated using a gradient descent algorithm with an L1 constraint, wherein the constraint strength increases linearly with the number of fault batches.

8. The full life cycle traceability digital shelf system of the Internet of Things according to claim 7 is characterized in that: The input of the long short-term memory network is multidimensional data including historical inventory, sales data and time series characteristics. The information is selectively memorized and forgotten through a gating mechanism to capture long-term dependencies in the data.

9. The full life cycle traceability digital shelf system of the Internet of Things according to claim 1 is characterized in that: When generating the heat map of the logistics trajectory for displaying consumer rights, the hierarchical data reconstruction unit adopts a kernel density estimation method to adaptively adjust the bandwidth according to the distribution of the logistics trajectory points, so as to intuitively display the density of the logistics trajectory.

10. The full life cycle traceability digital shelf system of the Internet of Things according to claim 1 is characterized in that: When the holographic traceability execution engine module receives a query request of the regulatory authority, in the process of calling the spatiotemporal anchor point account book to restore the equipment abnormality timeline, it combines the quantified indicators of the equipment vibration spectrum abnormality and the equipment health equation to determine the time point when the equipment abnormality occurs and the remaining service life assessment, and sets reasonable thresholds for key indicators. When the sensor data is abnormal, the system can immediately issue an early warning signal to remind relevant personnel to handle it in time; According to the data analysis results, based on equipment status data and product quality traceability information.

Images