Unattended warehouse intelligent monitoring system and method based on Internet of Things

Abstract

The invention discloses an unattended warehouse intelligent monitoring system and method based on the Internet of Things, and belongs to the technical field of intelligent monitoring of the Internet of Things. The system comprises a storage location management module, an RFID technology scanning module, an asset ex-warehouse analysis module, a data feedback module and a standby storage location analysis module. The output end of the storage location management module is connected with the input end of the RFID technology scanning module; the output end of the RFID technology scanning module is connected with the input end of the asset ex-warehouse analysis module; and the output end of the asset ex-warehouse analysis module is connected with the input end of the data feedback module. According to the invention, a set of unattended warehouse based on the RFID technology is established, an advanced, reliable and applicable digital platform for automatic identification and intelligent management of incoming and outgoing assets is provided, and the warehouse location in the warehouse is intelligently adjusted and processed, so that the real-time dynamic management capability of an enterprise on the internal assets is qualitatively improved.

Description

Technical Field

[0001] This invention relates to the field of IoT smart monitoring technology, specifically to an IoT-based unmanned warehouse smart monitoring system and method. Background Technology

[0002] In current enterprises and institutions, assets and equipment are diverse, numerous, and frequently used, leading to the following problems: potential risks of asset loss and lengthy safety inspections; inconsistent specifications and large quantities make it difficult to find tools; comprehensive inventory checks are required after use, but manual operations are inefficient and inventory is inaccurate; excessive maintenance time reduces work efficiency; some assets are used in harsh environments, making traditional barcode / QR code and visual methods ineffective for management; and there is a lack of asset status and real-time inventory monitoring systems, as well as a lack of data analysis and business forecasting. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent monitoring system and method for unmanned warehouses based on the Internet of Things, so as to solve the problems mentioned in the background art.

[0004] To address the aforementioned technical problems, this invention provides the following technical solution: a smart monitoring method for unmanned warehouses based on the Internet of Things, comprising the following steps:

[0005] S1. Construct an unmanned warehouse data system, which includes warehouse location management, RFID technology scanning, and backend management;

[0006] S2. Call the historical outbound asset information data in the unattended warehouse data system, build an asset outbound analysis model, and generate time analysis data for asset outbound and asset return.

[0007] S4. Use RFID technology to scan and obtain outbound asset data in real time, mark the storage location corresponding to the outbound asset. If the predicted outbound time of any storage location A is greater than the predicted return time of the marked storage location, then the asset in storage location A will be moved to the marked storage location. If the predicted outbound time of multiple storage locations is greater than the predicted return time of the marked storage location, then the maximum value among them will be selected as storage location A.

[0008] S4. Construct a time period T and record the number of backup management measures that occur within the time period T. The backup management measures refer to the management measures taken when assets are prematurely put into storage or assets in storage are not released in time due to the discrepancy between the predicted value and the actual value of the asset outbound analysis model. Construct a backup storage location analysis model, output the number of backup storage locations for the next time period T based on the backup storage location analysis model, and transmit it to the administrator port for review.

[0009] According to the above technical solution, the warehouse location management refers to the installation of RFID tags and indicator lights on each location in the unmanned warehouse, which automatically illuminate when assets are removed from the warehouse; the RFID scanning technology refers to the deployment of RFID readers at the warehouse entrances and exits, where the RFID readers monitor the RFID tags on the locations when assets are taken out of the warehouse, recording the asset removal record for administrator review; the backend management includes asset classification management and asset location management, classifying assets according to different usage characteristics, and simultaneously using RFID handheld devices to inventory asset information data and upload it to the server data terminal.

[0010] According to the above technical solution, the construction of the asset outbound analysis model includes:

[0011] S3-1. Classify assets (usually by unit, i.e., a certain group unit uses a certain type of asset, such as an environmental protection unit may use geographical assets. The overall classification should not be too many, and the specific classification depends on the amount of data). Obtain the historical outbound interval of any category of assets after classification, construct a training set of outbound interval of assets for that category, and construct a smoothing equation of outbound interval of assets for that category based on the training set of outbound interval of assets for that category.

[0012] S3-2, The smoothing equations include horizontal smoothing equations, trend smoothing equations, and seasonal smoothing equations:

[0013] U i =j1(x i -P i-t )+(1-k1)(U i-1 +I i-1 )

[0014] I i =k2(U i -U i-1 )+(1-k2)I i-1

[0015] P i =k3(x i -U i-1 -I i-1 )+(1-k3)P i-t

[0016] The predicted interval for the release of this type of asset is:

[0017] x i+c =U i +cI i +P i-t+1

[0018] Where i represents the current time interval data; t represents the step size of the training period; k1, k2, and k3 represent the smoothing parameters of the horizontal smoothing equation, the trend smoothing equation, and the seasonal smoothing equation, respectively; x i+c x represents the predicted value at the c-th interval; i Represents the actual value of the current period; U i I represents the current period's level smoothing forecast solution; i P represents the current period's level smoothing forecast solution; i This represents the current period's level smoothing forecast solution;

[0019] S3-3. Obtain the historical number of scrappings for this type of asset, calculate the average number of scrappings, and use it as the scrapping evaluation standard for this type of asset. If the current number of outbound shipments meets the scrapping evaluation standard for this type of asset, mark it as scrapped asset, and the return interval time value is set to infinity.

[0020] S3-4. If the current number of outbound shipments does not meet the evaluation criteria for the scrapping of this type of asset, take the historical interval data of asset return from the same period as S3-2, construct the same algorithm program as S3-2, and calculate and output the predicted value of the return interval time for this type of asset.

[0021] In the above technical solution, the sequence analysis equation is still used, and the time interval between outbound and return is based on the same historical data within the same period. The same algorithm is constructed to minimize the error impact between different algorithm programs. Specifically, the current period data x in the original smoothing equation is used... i Replace with return interval x j Based on the program, new smoothing parameters are given, and the same trend equation as S3-2 is constructed to generate new time points.

[0022] S3-5. Use RFID technology to scan and obtain outbound asset data in real time, and mark the storage location corresponding to the outbound asset. If the predicted outbound time of any storage location A is greater than the predicted return time of the marked storage location, the asset in storage location A will be moved to the marked storage location. If the predicted outbound time of multiple storage locations is greater than the predicted return time of the marked storage location, the maximum value among them will be selected as storage location A.

[0023] According to the above technical solution, the construction of the backup storage location analysis model includes:

[0024] Establish a time period T and record the number of backup management measures that occur within the time period T;

[0025] Obtain the number of backup management measures for each time period T under historical data, use the backup management measures in the historical data to train the neural network, and construct the input layer, hidden layer and output layer of the LSTM network;

[0026] The number of backup management measures for each time period T is recorded as the dataset. The dataset is randomly divided into training and validation sets according to a set ratio. The maximum number of training iterations K is set. After the maximum number of training iterations K is completed, it is used as the model output.

[0027] Construct a knowledge system, select the most recent number of backup management measures, and build an adaptive correction algorithm:

[0028]

[0029] Where R represents the number of the most recent backup management measures; R LSTM represents the predicted value output by the LSTM model after completing the maximum number of training iterations K; w represents the forgetting factor, which is greater than 0.5 and less than 1; M represents the number of backup storage locations in the predicted output, rounded up.

[0030] Because the training data mainly consists of error data from warehouse operations, a piecewise cross-validation method was used to enable the LSTM to learn more and avoid overfitting when the dataset size is small. For example, the dataset is randomly divided into a 7:3 ratio, with 70% of the data used as the training set and 30% as the validation set each time, until the neural network has been trained K times.

[0031] The knowledge system also aims to further improve the overall predictive feasibility by setting corresponding forgetting factors. The forgetting factor is a coefficient that essentially refers to the fact that as the number of iterations increases, old data accumulates more and more, overwhelming new data information and ultimately preventing parameter estimation and algorithm failure. Therefore, it is necessary to weight the old data during the iteration process to reduce its influence on the new data and increase its impact. This approach is applicable to this application, effectively enhancing the specific impact of new data and further improving the predictive accuracy.

[0032] The IoT-based unmanned warehouse intelligent monitoring system includes a warehouse location management module, an RFID technology scanning module, an asset outbound analysis module, a data feedback module, and a backup warehouse location analysis module.

[0033] The warehouse location management module installs RFID tags and indicator lights on each location in the unmanned warehouse, automatically illuminating when assets are removed. The RFID scanning module deploys RFID readers at warehouse entrances and exits; when assets are taken out of the warehouse, the readers monitor the RFID tags on the locations and record the asset removal information for administrator review. The asset removal analysis module retrieves historical asset removal information from the unmanned warehouse data system to build an asset removal analysis model and generate time-based analysis data for asset removal and return. The data feedback module uses RFID technology to scan and acquire asset removal data in real time, marking the corresponding warehouse location for each asset. If the predicted asset outbound time for any storage location A is greater than the asset return time for the marked storage location, the asset in storage location A will be moved to the marked storage location. If the predicted asset outbound time for multiple storage locations is greater than the asset return time for the marked storage location, the maximum value among them will be selected as storage location A. The standby storage location analysis module is used to construct a time period T and record the number of standby management measures that occur within the time period T. The standby management measures refer to the management measures taken when the predicted value of the asset outbound analysis model does not match the actual value, resulting in the early entry of outbound assets or the failure of assets in the storage to be outbound in a timely manner. The standby storage location analysis model is constructed, and based on the standby storage location analysis model, the number of standby storage locations for the next time period T is output and transmitted to the administrator port for review.

[0034] The output of the storage location management module is connected to the input of the RFID technology scanning module; the output of the RFID technology scanning module is connected to the input of the asset outbound analysis module; the output of the asset outbound analysis module is connected to the input of the data feedback module; and the output of the data feedback module is connected to the input of the backup storage location analysis module.

[0035] According to the above technical solution, the warehouse location management module includes an RFID tag placement unit and a light-up indicator unit;

[0036] The RFID tag deployment unit is used to install RFID tags on each storage location in the unattended warehouse; the light indicator unit is used to automatically light up to remind users when assets are taken out of the warehouse.

[0037] The output of the RFID tag placement unit is connected to the input of the light indicator unit.

[0038] According to the above technical solution, the RFID technology scanning module includes a smart monitoring unit and a recording unit;

[0039] The intelligent monitoring unit is used to deploy RFID readers at the warehouse entrances and exits. When an asset is taken out of the warehouse, the RFID reader will monitor the RFID tag on the storage location. The recording unit is used to record the asset transfer records for administrator review.

[0040] The output of the intelligent monitoring unit is connected to the input of the recording unit.

[0041] According to the above technical solution, the asset outbound analysis module includes a data retrieval unit and a model analysis unit;

[0042] The data retrieval unit is used to retrieve historical outbound asset information data from the unattended warehouse data system; the model analysis unit is used to construct an asset outbound analysis model based on the historical outbound asset information data, and generate time analysis data for asset outbound and asset return.

[0043] The output of the data retrieval unit is connected to the input of the model analysis unit.

[0044] According to the above technical solution, the data feedback module includes a storage location marking unit and a storage location adjustment unit;

[0045] The warehouse location marking unit uses RFID technology to scan and acquire outbound asset data in real time, marking the warehouse location corresponding to the outbound asset; the warehouse location adjustment unit is used to move the asset in warehouse location A to the marked warehouse location if the predicted outbound time of any warehouse location A is greater than the asset return time of the marked warehouse location. If the predicted outbound time of multiple warehouse locations is greater than the asset return time of the marked warehouse location, the maximum value among them is selected as warehouse location A.

[0046] The output of the storage location marking unit is connected to the input of the storage location adjustment unit.

[0047] According to the above technical solution, the backup storage location analysis module includes a backup management unit and a storage location quantity analysis unit;

[0048] The backup management unit is used to construct a time period T and record the number of backup management measures that occur within the time period T; the storage location quantity analysis unit is used to construct a backup storage location analysis model, output the number of backup storage locations for the next time period T based on the backup storage location analysis model, and transmit it to the administrator port for review.

[0049] The output of the backup management unit is connected to the input of the storage location quantity analysis unit.

[0050] Compared with existing technologies, the beneficial effects achieved by this invention are as follows: This invention establishes an unmanned warehouse based on RFID technology. The system uses RFID middleware as a platform and incorporates multiple processes such as warehousing, inventory, and outbound operations, enabling warehouse storage statistics and high-speed automatic recording of inbound and outbound transactions. It establishes a secure and reliable registered asset file using modern technological means, strengthens asset supervision through advanced technology, rationally allocates resources, reduces resource waste, and prevents asset loss. Based on actual conditions, it effectively solves the problems of chaotic and disorderly asset management and poor real-time performance. It provides an advanced, reliable, and applicable digital platform for automatic identification and intelligent management of incoming and outgoing assets, intelligently adjusting and processing warehouse locations, thus qualitatively improving the enterprise's ability to manage internal assets in real time. Different base stations and regional RFID readers can be added to automatically determine asset status (borrowed, transferred, idle, scrapped, etc.) using collected data. End users can perform asset data statistics and queries anytime, anywhere through a browser via a visual operating interface. Attached Figure Description

[0051] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0052] Figure 1 This is a schematic diagram of the intelligent monitoring system and method for unmanned warehouses based on the Internet of Things, according to the present invention. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] Please see Figure 1 In this first embodiment: an unmanned warehouse data system is constructed. Each storage location in the unmanned warehouse is equipped with an RFID tag and an indicator light. The light automatically illuminates when an asset is removed from the warehouse. The RFID scanning technology refers to the deployment of RFID readers at the warehouse entrances and exits. When an asset is taken out of the warehouse, the RFID reader monitors the RFID tag on the storage location and records the asset removal record for administrator review. The backend management includes asset classification management and asset storage location management. Assets are classified according to different usage characteristics, and asset information data is inventoried using RFID handheld devices and uploaded to the server data terminal.

[0055] S3-1. Classify the assets and obtain the historical outbound interval time of any category of assets after classification. Construct a training set of outbound interval time for that category of assets and construct a smoothing equation for that category of assets on the outbound interval time based on the training set of outbound interval time for that category of assets.

[0056] S3-2, The smoothing equations include horizontal smoothing equations, trend smoothing equations, and seasonal smoothing equations:

[0057] U i =k1(x i -P i-t )+(1-k1)(U i-1 +I i-1 )

[0058] I i =k2(U i -U i-1 )+(1-k2)I i-1

[0059] P i =k3(x i -U i-1 -I i-1 )+(1-k3)P i-t

[0060] The predicted interval for the release of this type of asset is:

[0061] x i+c =U i +cI i +P i-t+1

[0062] Where i represents the current time interval data; t represents the step size of the training period; k1, k2, and k3 represent the smoothing parameters of the horizontal smoothing equation, the trend smoothing equation, and the seasonal smoothing equation, respectively; x i+c x represents the predicted value at the c-th interval; i Represents the actual value of the current period; U i I represents the current period's level smoothing forecast solution; i P represents the current period's level smoothing forecast solution; i This represents the current period's level smoothing forecast solution;

[0063] S3-3. Obtain the historical number of scrappings for this type of asset, calculate the average number of scrappings, and use it as the scrapping evaluation standard for this type of asset. If the current number of outbound shipments meets the scrapping evaluation standard for this type of asset, mark it as scrapped asset, and the return interval time value is set to infinity.

[0064] S3-4. If the current number of outbound shipments does not meet the evaluation criteria for the scrapping of this type of asset, take the historical interval data of asset return from the same period as S3-2, construct the same algorithm program as S3-2, and calculate and output the predicted value of the return interval time for this type of asset.

[0065] S3-5. Use RFID technology to scan and obtain outbound asset data in real time, and mark the storage location corresponding to the outbound asset. If the predicted outbound time of any storage location A is greater than the predicted return time of the marked storage location, the asset in storage location A will be moved to the marked storage location. If the predicted outbound time of multiple storage locations is greater than the predicted return time of the marked storage location, the maximum value among them will be selected as storage location A.

[0066] The construction of the backup storage location analysis model includes:

[0067] Establish a time period T and record the number of backup management measures that occur within the time period T;

[0068] Obtain the number of backup management measures for each time period T under historical data, use the backup management measures in the historical data to train the neural network, and construct the input layer, hidden layer and output layer of the LSTM network;

[0069] The number of backup management measures for each time period T is recorded as the dataset. The dataset is randomly divided into training and validation sets according to a set ratio. The maximum number of training iterations K is set. After the maximum number of training iterations K is completed, it is used as the model output.

[0070] Construct a knowledge system, select the most recent number of backup management measures, and build an adaptive correction algorithm:

[0071]

[0072] Where R represents the number of the most recent backup management measures; R LSTM represents the predicted value output by the LSTM model after completing the maximum number of training iterations K; w represents the forgetting factor, which is greater than 0.5 and less than 1; M represents the number of backup storage locations in the predicted output, rounded up.

[0073] In this second embodiment, an IoT-based unmanned warehouse smart monitoring system is provided. The system includes a warehouse location management module, an RFID technology scanning module, an asset outbound analysis module, a data feedback module, and a backup warehouse location analysis module.

[0074] The warehouse location management module installs RFID tags and indicator lights on each location in the unmanned warehouse, automatically illuminating when assets are removed. The RFID scanning module deploys RFID readers at warehouse entrances and exits; when assets are taken out of the warehouse, the readers monitor the RFID tags on the locations and record the asset removal information for administrator review. The asset removal analysis module retrieves historical asset removal information from the unmanned warehouse data system to build an asset removal analysis model and generate time-based analysis data for asset removal and return. The data feedback module uses RFID technology to scan and acquire asset removal data in real time, marking the corresponding warehouse location for each asset. If the predicted asset outbound time for any storage location A is greater than the asset return time for the marked storage location, the asset in storage location A will be moved to the marked storage location. If the predicted asset outbound time for multiple storage locations is greater than the asset return time for the marked storage location, the maximum value among them will be selected as storage location A. The standby storage location analysis module is used to construct a time period T and record the number of standby management measures that occur within the time period T. The standby management measures refer to the management measures taken when the predicted value of the asset outbound analysis model does not match the actual value, resulting in the early entry of outbound assets or the failure of assets in the storage to be outbound in a timely manner. The standby storage location analysis model is constructed, and based on the standby storage location analysis model, the number of standby storage locations for the next time period T is output and transmitted to the administrator port for review.

[0075] The output of the storage location management module is connected to the input of the RFID technology scanning module; the output of the RFID technology scanning module is connected to the input of the asset outbound analysis module; the output of the asset outbound analysis module is connected to the input of the data feedback module; and the output of the data feedback module is connected to the input of the backup storage location analysis module.

[0076] The warehouse location management module includes an RFID tag placement unit and an indicator light unit;

[0077] The RFID tag deployment unit is used to install RFID tags on each storage location in the unattended warehouse; the light indicator unit is used to automatically light up to remind users when assets are taken out of the warehouse.

[0078] The output of the RFID tag placement unit is connected to the input of the light indicator unit.

[0079] The RFID technology scanning module includes a smart monitoring unit and a recording unit;

[0080] The intelligent monitoring unit is used to deploy RFID readers at the warehouse entrances and exits. When an asset is taken out of the warehouse, the RFID reader will monitor the RFID tag on the storage location. The recording unit is used to record the asset transfer records for administrator review.

[0081] The output of the intelligent monitoring unit is connected to the input of the recording unit.

[0082] The asset outbound analysis module includes a data retrieval unit and a model analysis unit;

[0083] The data retrieval unit is used to retrieve historical outbound asset information data from the unattended warehouse data system; the model analysis unit is used to construct an asset outbound analysis model based on the historical outbound asset information data, and generate time analysis data for asset outbound and asset return.

[0084] The output of the data retrieval unit is connected to the input of the model analysis unit.

[0085] The data feedback module includes a storage location marking unit and a storage location adjustment unit;

[0086] The warehouse location marking unit uses RFID technology to scan and acquire outbound asset data in real time, marking the warehouse location corresponding to the outbound asset; the warehouse location adjustment unit is used to move the asset in warehouse location A to the marked warehouse location if the predicted outbound time of any warehouse location A is greater than the asset return time of the marked warehouse location. If the predicted outbound time of multiple warehouse locations is greater than the asset return time of the marked warehouse location, the maximum value among them is selected as warehouse location A.

[0087] The output of the storage location marking unit is connected to the input of the storage location adjustment unit.

[0088] The backup storage location analysis module includes a backup management unit and a storage location quantity analysis unit;

[0089] The backup management unit is used to construct a time period T and record the number of backup management measures that occur within the time period T; the storage location quantity analysis unit is used to construct a backup storage location analysis model, output the number of backup storage locations for the next time period T based on the backup storage location analysis model, and transmit it to the administrator port for review.

[0090] The output of the backup management unit is connected to the input of the storage location quantity analysis unit.

[0091] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0092] Finally, it should be noted that the above descriptions are merely preferred embodiments of the present invention and are not intended to limit the present invention. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A smart monitoring method for unmanned warehouses based on the Internet of Things, characterized by: The method includes the following steps: S1. Construct an unmanned warehouse data system, which includes warehouse location management, RFID technology scanning, and backend management; S2. Call the historical outbound asset information data in the unattended warehouse data system, build an asset outbound analysis model, and generate time analysis data for asset outbound and asset return. S4. Use RFID technology to scan and obtain outbound asset data in real time, mark the storage location corresponding to the outbound asset. If the predicted outbound time of any storage location A is greater than the predicted return time of the marked storage location, then the asset in storage location A will be moved to the marked storage location. If the predicted outbound time of multiple storage locations is greater than the predicted return time of the marked storage location, then the maximum value among them will be selected as storage location A. S4. Construct a time period T and record the number of backup management measures that occur within the time period T. The backup management measures refer to the management measures taken when assets are prematurely put into storage or assets in storage are not released in time due to the discrepancy between the predicted value and the actual value of the asset outbound analysis model. Construct a backup storage location analysis model, output the number of backup storage locations for the next time period T based on the backup storage location analysis model, and transmit it to the administrator port for review.

2. The intelligent monitoring method for unmanned warehouses based on the Internet of Things according to claim 1, characterized in that: The warehouse location management refers to the installation of RFID tags and indicator lights on each location in the unmanned warehouse, which automatically illuminate when assets are removed from the warehouse. The RFID scanning technology refers to the deployment of RFID readers at the warehouse entrances and exits. When assets are taken out of the warehouse, the RFID readers monitor the RFID tags on the locations and record the asset removal records for administrator review. The backend management includes asset classification management and asset location management, which classifies assets according to different usage characteristics and uses RFID handheld devices to inventory asset information data and upload it to the server data terminal.

3. The intelligent monitoring method for unmanned warehouses based on the Internet of Things according to claim 2, characterized in that: The constructed asset outbound analysis model includes: S3-1. Classify the assets and obtain the historical outbound interval time of any category of assets after classification. Construct a training set of outbound interval time for that category of assets and construct a smoothing equation for that category of assets on the outbound interval time based on the training set of outbound interval time for that category of assets. S3-2, The smoothing equations include horizontal smoothing equations, trend smoothing equations, and seasonal smoothing equations: U i = k1(x i -P i-t )+(1 - k1)(U i-1 +I i-1 ) AND i =k2(U i -IN i-1 )+(1-k2)I i-1 P i =k3(x i -U i-1 -I i-1 )+(1-k3)P i-t The predicted interval for the release of this type of asset is: x i+c = U i +cI i +P i-t+1 Where i represents the current time interval data; t represents the step size of the training period; k1, k2, and k3 represent the smoothing parameters of the horizontal smoothing equation, the trend smoothing equation, and the seasonal smoothing equation, respectively; x i+c x represents the predicted value at the c-th interval; i U represents the actual value for the current period; i I represents the current period's level smoothing forecast solution; i P represents the current period's level smoothing forecast solution; i This represents the current period's level smoothing forecast solution; S3-3. Obtain the historical number of scrappings for this type of asset, calculate the average number of scrappings, and use it as the evaluation standard for scrapping this type of asset. If the current number of outbound shipments meets the evaluation standard for scrapping this type of asset, mark it as scrapped asset, and the return interval time value is set to infinity. S3-4. If the current number of outbound shipments does not meet the evaluation criteria for the scrapping of this type of asset, take the historical interval data of asset return from the same period as S3-2, construct the same algorithm program as S3-2, and calculate and output the predicted value of the return interval time for this type of asset. S3-5. Use RFID technology to scan and obtain outbound asset data in real time, and mark the storage location corresponding to the outbound asset. If the predicted outbound time of any storage location A is greater than the predicted return time of the marked storage location, the asset in storage location A will be moved to the marked storage location. If the predicted outbound time of multiple storage locations is greater than the predicted return time of the marked storage location, the maximum value among them will be selected as storage location A.

4. The intelligent monitoring method for unmanned warehouses based on the Internet of Things according to claim 3, characterized in that: The construction of the backup storage location analysis model includes: Establish a time period T and record the number of backup management measures that occur within the time period T; Obtain the number of backup management measures for each time period T under historical data, use the backup management measures in the historical data to train the neural network, and construct the input layer, hidden layer and output layer of the LSTM network; The number of backup management measures for each time period T is recorded as the dataset. The dataset is randomly divided into training and validation sets according to a set ratio. The maximum number of training iterations K is set. After the maximum number of training iterations K is completed, it is used as the model output. Construct a knowledge system, select the most recent number of backup management measures, and build an adaptive correction algorithm: Where R represents the number of the most recent backup management measures; R LSTM represents the predicted value output by the LSTM model after completing the maximum number of training iterations K; w represents the forgetting factor, which is greater than 0.5 and less than 1; M represents the number of backup storage locations in the predicted output, rounded up.

5. An IoT-based intelligent monitoring system for unmanned warehouses, characterized in that: The system includes a warehouse location management module, an RFID technology scanning module, an asset outbound analysis module, a data feedback module, and a backup warehouse location analysis module; The warehouse location management module installs RFID tags and indicator lights on each location in the unmanned warehouse, automatically illuminating when assets are removed. The RFID scanning module deploys RFID readers at warehouse entrances and exits; when assets are taken out of the warehouse, the readers monitor the RFID tags on the locations and record the asset removal information for administrator review. The asset removal analysis module retrieves historical asset removal information from the unmanned warehouse data system to build an asset removal analysis model and generate time-based analysis data for asset removal and return. The data feedback module uses RFID technology to scan and acquire asset removal data in real time, marking the corresponding warehouse location for each asset. If the predicted asset outbound time for any storage location A is greater than the asset return time for the marked storage location, the asset in storage location A will be moved to the marked storage location. If the predicted asset outbound time for multiple storage locations is greater than the asset return time for the marked storage location, the maximum value among them will be selected as storage location A. The standby storage location analysis module is used to construct a time period T and record the number of standby management measures that occur within the time period T. The standby management measures refer to the management measures taken when the predicted value of the asset outbound analysis model does not match the actual value, resulting in the early entry of outbound assets or the failure of assets in the storage to be outbound in a timely manner. The standby storage location analysis model is constructed, and based on the standby storage location analysis model, the number of standby storage locations for the next time period T is output and transmitted to the administrator port for review. The output of the storage location management module is connected to the input of the RFID technology scanning module; the output of the RFID technology scanning module is connected to the input of the asset outbound analysis module; the output of the asset outbound analysis module is connected to the input of the data feedback module; and the output of the data feedback module is connected to the input of the backup storage location analysis module.

6. The IoT-based intelligent monitoring system for unmanned warehouses according to claim 5, characterized in that: The warehouse location management module includes an RFID tag placement unit and an indicator light unit; The RFID tag deployment unit is used to install RFID tags on each storage location in the unattended warehouse; the light indicator unit is used to automatically light up to remind users when assets are taken out of the warehouse. The output of the RFID tag placement unit is connected to the input of the light indicator unit.

7. The IoT-based intelligent monitoring system for unmanned warehouses according to claim 5, characterized in that: The RFID technology scanning module includes a smart monitoring unit and a recording unit; The intelligent monitoring unit is used to deploy RFID readers at the warehouse entrances and exits. When an asset is taken out of the warehouse, the RFID reader will monitor the RFID tag on the storage location. The recording unit is used to record the asset transfer records for administrator review. The output of the intelligent monitoring unit is connected to the input of the recording unit.

8. The IoT-based intelligent monitoring system for unmanned warehouses according to claim 5, characterized in that: The asset outbound analysis module includes a data retrieval unit and a model analysis unit; The data retrieval unit is used to retrieve historical outbound asset information data from the unattended warehouse data system; the model analysis unit is used to construct an asset outbound analysis model based on the historical outbound asset information data, and generate time analysis data for asset outbound and asset return. The output of the data retrieval unit is connected to the input of the model analysis unit.

9. The IoT-based intelligent monitoring system for unmanned warehouses according to claim 5, characterized in that: The data feedback module includes a storage location marking unit and a storage location adjustment unit; The warehouse location marking unit uses RFID technology to scan and acquire outbound asset data in real time, marking the warehouse location corresponding to the outbound asset; the warehouse location adjustment unit is used to move the asset in warehouse location A to the marked warehouse location if the predicted outbound time of any warehouse location A is greater than the asset return time of the marked warehouse location. If the predicted outbound time of multiple warehouse locations is greater than the asset return time of the marked warehouse location, the maximum value among them is selected as warehouse location A. The output of the storage location marking unit is connected to the input of the storage location adjustment unit.

10. The IoT-based intelligent monitoring system for unmanned warehouses according to claim 5, characterized in that: The backup storage location analysis module includes a backup management unit and a storage location quantity analysis unit; The backup management unit is used to construct a time period T and record the number of backup management measures that occur within the time period T; the storage location quantity analysis unit is used to construct a backup storage location analysis model, output the number of backup storage locations for the next time period T based on the backup storage location analysis model, and transmit it to the administrator port for review. The output of the backup management unit is connected to the input of the storage location quantity analysis unit.

Images