A logistics sorting system for cross-border e-commerce

Abstract

This invention discloses a logistics sorting system for cross-border e-commerce, comprising: a reverse logistics preprocessing module integrating an RFID reader and an X-ray image acquisition device, which performs quality inspection image acquisition on returned packages entering the sorting line and generates a returned product quality classification code; a graded diversion control module, which diverts returned packages to the forward sorting main line or the reverse buffer branch line according to the quality code; and a dynamic virtual grid sorting module, which, upon receiving a fragmented order containing multiple SKUs, identifies multiple storage location coordinates corresponding to multiple SKUs under the fragmented order and assigns a logical group ID to the fragmented order, and generates asynchronous handling instructions by calculating the shortest convergence time of multiple SKUs to the merging buffer area, triggering logistics product route matching and export customs declaration matching only when all SKUs associated with the logical group ID are gathered in the merging buffer area.

Description

Technical Field

[0001] This invention relates to the field of logistics sorting technology, and in particular to a logistics sorting system for cross-border e-commerce. Background Technology

[0002] Intelligent logistics sorting systems for cross-border e-commerce integrate cutting-edge technologies such as artificial intelligence, machine vision, the Internet of Things, and big data to build a fully automated solution covering the entire process from goods warehousing, information collection, intelligent identification, automatic classification, and route planning to precise sorting, packaging and labeling, and outbound loading. This achieves an exponential increase in sorting efficiency and extreme optimization of sorting accuracy. At the same time, dynamic route planning and resource scheduling algorithms reduce equipment energy consumption and space occupation, while real-time data monitoring and anomaly early warning mechanisms ensure the transparency and controllability of the sorting process. Ultimately, this achieves the goals of shortening order fulfillment cycles, reducing logistics operating costs, and improving customer experience, driving the entire cross-border e-commerce logistics industry towards intelligent, standardized, and green transformation and upgrading. It provides key technical support for cross-border e-commerce enterprises to build global supply chain competitiveness, helping them seize opportunities in fierce market competition and achieve sustainable development.

[0003] In existing technologies, returned goods need to be re-inspected, classified, and stored. Traditional sorting systems are mostly designed for forward logistics, resulting in a large number of returned goods that need to be verified, relabeled, or repaired being mixed with forward goods awaiting shipment at the entrance of the sorting main line. The processing time for returns accounts for more than 40% of the overall fulfillment cycle, which in turn causes wave execution blockages and sorting line downtime due to the excessively high proportion of time spent processing returned goods on the forward sorting main line. Therefore, a logistics sorting system for cross-border e-commerce is proposed. Summary of the Invention

[0004] The purpose of this invention is to address the shortcomings of existing technologies by proposing a logistics sorting system for cross-border e-commerce.

[0005] To achieve the above objectives, the present invention adopts the following technical solution: A logistics sorting system for cross-border e-commerce includes: The reverse logistics preprocessing module is located at the front end of the system parameter entry. The reverse logistics preprocessing module integrates an RFID reading device and an X-ray image acquisition device. It is used to perform quality inspection image acquisition on returned packages entering the sorting line and generate corresponding returned condition grading codes. The returned condition grading codes include condition codes that characterize the physical state of the goods and near-expiry coefficients for beauty products. The graded diversion control module is communicatively connected to the reverse logistics preprocessing module. The graded diversion control module diverts returned packages to the forward sorting main line or the reverse buffer branch line according to the condition code. Among them, the resaleable goods determined to be of category A are sent to the forward sorting main line and processed according to the preset destination country matching logic. The goods determined to be of category B or C that need to be processed are intercepted to the reverse buffer branch line for batch waiting. The dynamic virtual grid sorting module is connected to the forward sorting main line. The dynamic virtual grid sorting module has a built-in real-time order tree decomposition algorithm, which is used to identify multiple storage location coordinates corresponding to multiple SKUs under the fragmented order when receiving a fragmented order containing multiple inventory units (SKUs). It assigns a logical group ID to the fragmented order instead of a physical grid. It generates asynchronous handling instructions by calculating the shortest convergence time of multiple SKUs to the merging buffer area. Only when all SKUs associated with the logical group ID are gathered in the merging buffer area will the logistics product route matching and export customs declaration matching process be triggered. The dual-modal data verification and correction module is connected to the reverse logistics preprocessing module and the forward sorting mainline, respectively. The dual-modal data verification and correction module has a built-in weight-volume density verification model and a self-healing address tree matching model. When the barcode of a returned package is damaged or the address information is missing, it automatically triggers weighing and 3D visual measurement to correct the product name by reverse matching the SKU-ID in the database through the measured density, or it traces back the recipient's historical logistics trajectory heat map to obtain the most frequently occurring geographical coordinates as the default sorting basis.

[0006] The above technical solution further includes: Furthermore, the quality code in the reverse logistics preprocessing module includes grades A, B, and C. Grade A represents goods that are in perfect condition in terms of appearance and contents and can be directly entered into the forward sorting line for resale. Grade B represents goods that are in good condition but have defects in packaging or labels and need to be relabeled before they can be put back into the warehouse. Grade C represents goods that are damaged, have liquid leakage, or have functional failures and need to be transferred to the repair process or disposed of. The graded diversion control module triggers different transmission path guidance instructions according to the judgment results of grades A, B, and C.

[0007] Furthermore, the reverse logistics preprocessing module includes a cosmetics shelf-life calculation submodule. This submodule obtains the production batch number information of the goods in the returned package through the RFID reading device and automatically calculates the near-expiration coefficient in combination with the current system processing date. The near-expiration coefficient is defined as the percentage of the remaining valid shelf life of the goods relative to the total shelf life specified by the manufacturer. When the near-expiration coefficient is lower than a preset threshold, the graded diversion control module forcibly covers or downgrades the condition code of the returned package to category B or C to prevent near-expiration goods from being mixed into the forward sorting mainline.

[0008] Furthermore, when a single order received by the system contains 5 or more SKUs of inventory units, and the volume of goods corresponding to a single SKU of inventory unit in the order is less than or equal to 8000 cubic centimeters, the dynamic virtual grid sorting module stops calling the processing logic of the fixed wave aggregation strategy, and establishes a dynamic order tree structure diagram for the order. The data structure of the dynamic order tree structure diagram includes the set of three-dimensional storage location coordinates of each SKU of inventory unit in the order item in the storage space, as well as the set of conveyor path topology distances from each three-dimensional storage location coordinate to the merging buffer area.

[0009] Furthermore, the dynamic virtual grid sorting module generates an asynchronous transport instruction set for a single logical group ID. Each asynchronous transport instruction in the set is sent to the control unit of an autonomous mobile robot or a multi-layer shuttle conveyor line. Each asynchronous transport instruction includes the three-dimensional storage location coordinates of the corresponding inventory unit SKU, as well as the latest departure timest and the latest arrival timest of the instruction. The difference between the latest departure timest and the latest arrival timest is greater than or equal to the quotient obtained by dividing the conveyor path topology distance corresponding to the inventory unit SKU by the rated operating speed of the autonomous mobile robot or the multi-layer shuttle conveyor line. When the number of inventory unit SKU barcodes continuously read by the barcode verification device set at the entrance of the merging buffer area within the preset order set timeout threshold is equal to the total number of inventory unit SKUs associated with the logical group ID recorded in the dynamic order tree structure diagram, the dynamic virtual grid sorting module sends a release signal for the logical group ID to the logistics product route matching interface.

[0010] Furthermore, the weight-volume density verification model in the dual-modal data verification and correction module is configured with a density tolerance range database corresponding to irregularly shaped clothing soft bags and irregularly shaped container cosmetics in the fashion and beauty category. When the RFID reading device set on the conveyor belt fails to read the one-dimensional or two-dimensional barcode on the surface of the same returned package more than or equal to 3 times, and the returned package stays on the conveyor belt for more than 500 milliseconds, the dual-modal data verification and correction module sends a forced trigger acquisition command to the dynamic weighing sensor located below the conveyor belt and the 3D structured light vision camera located above the conveyor belt. The dual-modal data verification and correction module takes the ratio of the received measured weight value to the measured volume value as the measured density value, and compares the measured density value with each standard density range in the density tolerance range database. When there is any standard density range that makes the measured density value fall within the standard density range, the dual-modal data verification and correction module writes the SKU-ID corresponding to the standard density range into the system record field of the returned package.

[0011] Furthermore, the self-healing address tree matching model is connected to a historical order trajectory database, which stores the last-mile delivery geographic coordinates of all successfully delivered orders within a preset time period for a specific recipient's name and contact number combination. When parsing the destination country matching partition field corresponding to a returned package or a forward order, if the postal code field is missing or the detailed address field is unstructured natural language and cannot be matched with the standard address database, the self-healing address tree matching model retrieves the most frequent historical delivery coordinates associated with the recipient's name and contact number combination in the historical order trajectory database and temporarily assigns the most frequent historical delivery coordinates to the sorting destination parameter of the current task, thereby bypassing the manual address verification node to maintain the continuous flow of the sorting mainline.

[0012] Furthermore, a high-speed cache configuration module is included. This module uses high-frequency memory as a hardware carrier to load and store in real time the judgment rule base of the phase code, the threshold parameter table of the near-expiration coefficient, the logical group ID state machine generated by the real-time order tree decomposition algorithm, and the standard density interval data required by the dual-modal data verification and correction module. This enables each module to read the configuration data and intermediate state data with nanosecond-level low latency during the sorting process, ensuring that the quality inspection image acquisition, asynchronous handling instruction issuance, and self-healing address matching query process are synchronized with the running rhythm of the forward sorting mainline.

[0013] The present invention has the following beneficial effects: In this invention, by setting up a reverse logistics preprocessing module and a graded diversion control module, the returned goods of categories B / C that need to be processed are diverted to the reverse buffer branch line, and the return processing process is separated from the high-speed forward sorting main line, thus eliminating the wave blockage and downtime waiting of the forward sorting main line caused by repeated verification and relabeling of returned goods. Attached Figure Description

[0014] Figure 1 This is a system block diagram of a logistics sorting system for cross-border e-commerce proposed in this invention. Detailed Implementation

[0015] 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.

[0016] Please see Figure 1 As shown, the present invention is a logistics sorting system for cross-border e-commerce, comprising: The reverse logistics preprocessing module is located at the front end of the system parameter entry. The reverse logistics preprocessing module integrates an RFID reading device and an X-ray image acquisition device. It is used to perform quality inspection image acquisition on returned packages entering the sorting line and generate corresponding returned condition grading codes. The returned condition grading codes include condition codes that characterize the physical state of the goods and near-expiry coefficients for beauty products. The graded diversion control module is communicatively connected to the reverse logistics preprocessing module. The graded diversion control module diverts returned packages to the forward sorting main line or the reverse buffer branch line according to the condition code. Among them, the resaleable goods determined to be of category A are sent to the forward sorting main line and processed according to the preset destination country matching logic. The goods determined to be of category B or C that need to be processed are intercepted to the reverse buffer branch line for batch waiting. The dynamic virtual grid sorting module is connected to the forward sorting main line. The dynamic virtual grid sorting module has a built-in real-time order tree decomposition algorithm, which is used to identify multiple storage location coordinates corresponding to multiple SKUs under the fragmented order when receiving a fragmented order containing multiple inventory units (SKUs). It assigns a logical group ID to the fragmented order instead of a physical grid. It generates asynchronous handling instructions by calculating the shortest convergence time of multiple SKUs to the merging buffer area. Only when all SKUs associated with the logical group ID are gathered in the merging buffer area will the logistics product route matching and export customs declaration matching process be triggered. The dual-modal data verification and correction module is connected to the reverse logistics preprocessing module and the forward sorting mainline, respectively. The dual-modal data verification and correction module has a built-in weight-volume density verification model and a self-healing address tree matching model. When the barcode of a returned package is damaged or the address information is missing, it automatically triggers weighing and 3D visual measurement to correct the product name by reverse matching the SKU-ID in the database through the measured density, or it traces back the recipient's historical logistics trajectory heat map to obtain the most frequently occurring geographical coordinates as the default sorting basis.

[0017] In one embodiment, the quality code in the reverse logistics preprocessing module includes grades A, B, and C. Grade A represents goods that are in perfect condition and can be directly entered into the forward sorting line for resale. Grade B represents goods that are in good condition but have defects in packaging or labels and require simple relabeling before they can be put back into the warehouse. Grade C represents goods that are damaged, have liquid leakage, or have functional failures and need to be transferred to the repair process or disposed of. The graded diversion control module triggers different transmission path guidance instructions according to the determination results of grades A, B, and C.

[0018] In one embodiment, the reverse logistics preprocessing module includes a cosmetics shelf-life calculation submodule. This submodule obtains the production batch number information of the goods in the returned package through the RFID reading device and automatically calculates the near-expiration coefficient in combination with the current system processing date. The near-expiration coefficient is defined as the percentage of the remaining valid shelf life of the goods relative to the total shelf life specified by the manufacturer. When the near-expiration coefficient is lower than a preset threshold, the graded diversion control module forcibly covers or downgrades the condition code of the returned package to category B or C for processing, in order to prevent near-expiration goods from being mixed into the forward sorting mainline.

[0019] In one embodiment, when a single order received by the system contains 5 or more SKUs and the volume of goods corresponding to a single SKU in the order is less than or equal to 8000 cubic centimeters, the dynamic virtual grid sorting module stops calling the processing logic of the fixed wave aggregation strategy and establishes a separate dynamic order tree structure diagram for the order. The data structure of the dynamic order tree structure diagram includes the set of three-dimensional storage location coordinates of each SKU under the order in the storage space and the set of conveyor path topology distances from each three-dimensional storage location coordinate to the merging buffer area.

[0020] In one embodiment, the dynamic virtual grid sorting module generates a set of asynchronous transport instructions for a single logical group ID. Each asynchronous transport instruction in the set is sent to the control unit of an autonomous mobile robot or a multi-layer shuttle conveyor. Each asynchronous transport instruction includes the three-dimensional storage location coordinates of the corresponding inventory unit SKU, as well as the latest departure timest and the latest arrival timest of the instruction. The difference between the latest departure timest and the latest arrival timest is greater than or equal to the quotient obtained by dividing the conveyor path topology distance corresponding to the inventory unit SKU by the rated operating speed of the autonomous mobile robot or the multi-layer shuttle conveyor. When the number of inventory unit SKU barcodes continuously read by the barcode verification device set at the entrance of the merging buffer area within the preset order timeout threshold is equal to the total number of inventory unit SKUs associated with the logical group ID recorded in the dynamic order tree structure diagram, the dynamic virtual grid sorting module sends a release signal for the logical group ID to the logistics product route matching interface.

[0021] In one embodiment, the weight-volume density verification model in the dual-modal data verification and correction module is configured with a density tolerance range database corresponding to irregularly shaped clothing soft bags and irregularly shaped container cosmetics in the fashion and beauty category. When the RFID reader set on the conveyor belt fails to read the one-dimensional or two-dimensional barcode on the surface of the same returned package three or more times in a row, and the returned package stays on the conveyor belt for more than 500 milliseconds, the dual-modal data verification and correction module sends a forced trigger acquisition command to the dynamic weighing sensor located below the conveyor belt and the 3D structured light vision camera located above the conveyor belt. The dual-modal data verification and correction module takes the ratio of the received measured weight value to the measured volume value as the measured density value, and compares the measured density value with each standard density range in the density tolerance range database. When there is any standard density range that makes the measured density value fall within the standard density range, the dual-modal data verification and correction module writes the SKU-ID corresponding to the standard density range into the system record field of the returned package.

[0022] In one embodiment, the self-healing address tree matching model is connected to a historical order trajectory database. The historical order trajectory database stores the last-mile delivery geographic coordinates of all successfully delivered orders within a preset time period for a specific recipient's name and contact number combination. When parsing the destination country matching partition field corresponding to a returned package or a forward order, if the postal code field is missing or the detailed address field is unstructured natural language and cannot be matched with the standard address database, the self-healing address tree matching model retrieves the most frequent historical delivery coordinates associated with the recipient's name and contact number combination in the historical order trajectory database and temporarily assigns the most frequent historical delivery coordinates to the sorting destination parameter of the current task, thereby bypassing the manual address verification node to maintain the continuous flow of the sorting mainline.

[0023] In one embodiment, a high-speed cache configuration module is included. The high-speed cache configuration module uses high-frequency memory as a hardware carrier to load and store in real time the judgment rule base of the phase code, the threshold parameter table of the near-expiration coefficient, the logical group ID state machine generated by the real-time order tree decomposition algorithm, and the standard density interval data required by the dual-modal data verification and correction module. This enables each module to read the configuration data and intermediate state data with nanosecond-level low latency during the sorting process, ensuring that the quality inspection image acquisition, asynchronous handling instruction issuance, and self-healing address matching query process are synchronized with the running rhythm of the forward sorting mainline.

[0024] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A logistics sorting system for cross-border e-commerce, characterized in that, include: Reverse logistics preprocessing module: It integrates an RFID reading device and an X-ray image acquisition device, which are used to perform quality inspection image acquisition on returned packages entering the sorting line and generate a returned condition grading code. The returned condition grading code includes a condition code that represents the physical state of the goods and a near-expiry coefficient for cosmetic products. Graded diversion control module: According to the condition code, the returned packages are diverted to the forward sorting main line or the reverse buffer branch line. Among them, the goods determined to be of category A are sent to the forward sorting main line and processed according to the preset destination country matching logic. The goods determined to be category B or C are intercepted and sent to the reverse buffer branch line for batch waiting. Dynamic virtual grid sorting module: Built-in real-time order tree decomposition algorithm, used to identify multiple storage location coordinates corresponding to multiple SKUs under the fragmented order when receiving fragmented orders containing multiple inventory units (SKUs), and assign a logical group ID to the fragmented order. By calculating the shortest convergence time of multiple SKUs to the merging buffer area, asynchronous handling instructions are generated. Only when all SKUs associated with the logical group ID are gathered in the merging buffer area, logistics product route matching and export customs declaration matching are triggered. Dual-modal data verification and correction module: Built-in weight-volume density verification model and self-healing address tree matching model, used to automatically trigger weighing and 3D visual measurement when the barcode of the returned package is damaged or the address information is missing, so as to correct the product name by back matching the SKU-ID in the database through the measured density, or to trace back the recipient's historical logistics trajectory heat map to obtain the most frequently occurring geographical coordinates as the default sorting basis.

2. The logistics sorting system for cross-border e-commerce according to claim 1, characterized in that, The quality code in the reverse logistics preprocessing module includes grades A, B, and C. The graded diversion control module triggers different transmission path guidance instructions according to the determination results of grades A, B, and C.

3. A logistics sorting system for cross-border e-commerce according to claim 1, characterized in that, The reverse logistics preprocessing module includes a cosmetics shelf-life calculation submodule. The cosmetics shelf-life calculation submodule obtains the production batch number information of the goods in the returned package through the RFID reading device, and automatically calculates the near-expiration coefficient in combination with the current system processing date. When the near-expiration coefficient is lower than a preset threshold, the graded diversion control module forcibly covers or downgrades the condition code of the returned package to category B or category C for processing.

4. A logistics sorting system for cross-border e-commerce according to claim 1, characterized in that, When a single order received by the system contains 5 or more SKUs and the volume of goods corresponding to a single SKU in the order is less than or equal to 8000 cubic centimeters, the dynamic virtual grid sorting module stops calling the processing logic of the fixed wave aggregation strategy and establishes a dynamic order tree structure diagram for the order. The data structure of the dynamic order tree structure diagram includes the set of three-dimensional storage location coordinates of each SKU under the order in the storage space and the set of conveyor path topology distances from each three-dimensional storage location coordinate to the merging buffer area.

5. A logistics sorting system for cross-border e-commerce according to claim 4, characterized in that, The dynamic virtual grid sorting module generates an asynchronous transport instruction set for a single logical group ID. Each asynchronous transport instruction in the set is sent to the control unit of an autonomous mobile robot or a multi-layer shuttle conveyor. Each asynchronous transport instruction includes the three-dimensional storage location coordinates of the corresponding inventory unit SKU, as well as the latest departure timest and latest arrival timest of the instruction. The difference between the latest departure timest and the latest arrival timest is greater than or equal to the quotient obtained by dividing the conveyor path topology distance corresponding to the inventory unit SKU by the rated operating speed of the autonomous mobile robot or the multi-layer shuttle conveyor. When the number of inventory unit SKU barcodes continuously read by the barcode verification device set at the entrance of the merging buffer area within the preset order collection timeout threshold is equal to the total number of inventory unit SKUs associated with the logical group ID recorded in the dynamic order tree structure diagram, the dynamic virtual grid sorting module sends a release signal for the logical group ID to the logistics product route matching interface.

6. A logistics sorting system for cross-border e-commerce according to claim 1, characterized in that, The weight-volume density verification model in the dual-modal data verification and correction module is configured with a density tolerance interval database. When the RFID reader on the conveyor belt fails to read the one-dimensional or two-dimensional barcode on the surface of the same returned package three or more times in a row, and the returned package stays on the conveyor belt for more than 500 milliseconds, the dual-modal data verification and correction module sends a forced trigger acquisition command to the dynamic weighing sensor located below the conveyor belt and the 3D structured light vision camera located above the conveyor belt. The dual-modal data verification and correction module takes the ratio of the received measured weight value to the measured volume value as the measured density value, and compares the measured density value with each standard density interval in the density tolerance interval database. When there is any standard density interval that makes the measured density value fall within the standard density interval, the dual-modal data verification and correction module writes the SKU-ID corresponding to the standard density interval into the record field of the returned package.

7. A logistics sorting system for cross-border e-commerce according to claim 1, characterized in that, The self-healing address tree matching model is connected to a historical order trajectory database, which stores the last-mile delivery geographic coordinates of all successfully delivered orders within a preset time period for a specific recipient's name and contact number combination. When parsing the destination country matching partition field corresponding to a returned package or a forward order, if the postal code field is missing or the detailed address field is unstructured natural language and cannot be matched with the standard address database, the self-healing address tree matching model retrieves the most frequent historical delivery coordinates associated with the recipient's name and contact number combination in the historical order trajectory database and temporarily assigns the most frequent historical delivery coordinates to the sorting destination parameter of the current task.

8. A logistics sorting system for cross-border e-commerce according to claim 1, characterized in that, It includes a high-speed cache configuration module, which is used to load and store in real time the judgment rule base of the phase code, the threshold parameter table of the near-expiration coefficient, the logical group ID state machine generated by the real-time order tree decomposition algorithm, and the standard density interval data required by the dual-modal data verification and correction module.

Images