Intelligent sorting scheduling method and system based on multi-node information cooperation

Abstract

The application relates to the technical field of intelligent scheduling and automatic sorting of production logistics, and discloses an intelligent sorting scheduling method and system based on multi-node information cooperation, which provides a closed-loop mechanism of multi-node information cooperation-global state sensing-conflict prediction-dynamic scheduling, realizes timestamp synchronization and feature alignment of multi-node data through an edge cooperative gateway, realizes intelligent scheduling of sorting tasks and cooperative control of execution equipment through global state graph maintenance, path conflict prediction model and dynamic priority scheduling algorithm of a central intelligent scheduling engine, and solves core technical problems such as scheduling delay, frequent sorting path conflicts and low equipment resource utilization caused by multi-node information islands in traditional production logistics sorting scenes, so that the sorting scheduling process is automated, intelligentized and efficient.

Description

Technical Field

[0001] This invention relates to the field of intelligent scheduling and automated sorting technology in production logistics, specifically to an intelligent sorting and scheduling method and system based on multi-node information collaboration. Background Technology

[0002] Production logistics is a key link connecting manufacturing, warehousing and distribution. Among them, sorting and scheduling is the core operation to achieve efficient flow of materials from the input end to the target output end. As the manufacturing industry transforms and upgrades towards intelligence and flexibility, the types of materials in production logistics scenarios are increasing, sorting paths are becoming more complex, and the requirements for response timeliness are constantly increasing. Traditional sorting and scheduling methods can no longer meet the high-efficiency, high-precision and low-cost operation needs of modern production logistics.

[0003] In existing solutions, sorting and scheduling operations in most small and medium-sized manufacturing plants and warehousing and logistics centers still rely on manual operation. However, the manual sorting and scheduling model has the following significant drawbacks:

[0004] The scheduling efficiency is low. Manual scheduling relies on operators' memory and experience in judging material information (type, destination, urgency). When the sorting workload increases and the types of materials increase, it is difficult for operators to make the optimal scheduling decision in a short period of time.

[0005] The problem of information silos among multiple nodes leads to scheduling delays. In production logistics scenarios, the lack of a unified coordination and alignment mechanism for multi-source data such as RFID, vision sensors, position encoders, and sorting port status sensors in time and space results in delayed scheduling decisions and an inability to meet the timeliness requirements of real-time sorting.

[0006] The problem of frequent sorting path conflicts arises when multiple materials flow simultaneously on multiple conveyor belts. When the target sorting ports of different materials intersect or compete, the existing solutions lack effective conflict prediction and avoidance mechanisms, leading to material collisions, missorting, and system downtime.

[0007] The problem of low resource utilization of sorting equipment is that the load distribution of various sorting ports, AGVs, robotic arms and other execution equipment in the existing solution is uneven. Some equipment is running at full load for a long time while other equipment is idle, resulting in low overall resource utilization.

[0008] The scheduling strategy lacks dynamic adaptability. The existing solution uses static sorting rules, which cannot be dynamically adjusted according to real-time operating conditions (sorting port congestion, equipment failure, material surge), resulting in poor system robustness. Summary of the Invention

[0009] The purpose of this invention is to provide an intelligent sorting and scheduling method and system based on multi-node information collaboration to solve the problems mentioned in the background art.

[0010] To achieve the above objectives, the present invention provides the following technical solution:

[0011] An intelligent sorting and scheduling method based on multi-node information collaboration includes:

[0012] The sensor data of multi-source information nodes is synchronized through the edge collaboration gateway, and the synchronized multi-source sensor data is preprocessed at the edge to obtain standardized collaborative data packets.

[0013] A global state diagram of the sorting system's real-time status is constructed through a central intelligent scheduling engine, and the state vectors and attribute vectors of each material in the global state diagram are updated based on a collaborative data table.

[0014] Based on the updated global state diagram, the path of materials in transit is simulated to determine whether a path conflict has occurred and to calculate the corresponding conflict severity index.

[0015] Priority rescheduling is performed based on the conflict severity index and the load distribution of each sorting station in order to obtain and establish the optimal load-balanced scheduling scheme.

[0016] The optimal scheduling scheme is converted into scheduling instructions and sent to the execution device through the edge collaboration gateway to control the execution device to perform sorting actions;

[0017] The execution results are fed back through the edge collaboration gateway, and the global state graph is updated based on the execution results.

[0018] As a further aspect of the present invention: the step of synchronizing sensor data of multi-source information nodes through an edge collaborative gateway, and performing edge preprocessing on the synchronized multi-source sensor data to obtain standardized collaborative data packets specifically includes:

[0019] Sensor data from multiple sensor nodes, including RFID tag reading nodes, visual inspection nodes, conveyor belt position sensor nodes, and sorting port load status sensor nodes, are synchronized through an edge collaboration gateway.

[0020] Based on the clock reference and time synchronization protocol of the central intelligent scheduling engine, the timestamps of several sensor nodes associated with the edge collaborative gateway are synchronized, and the sensor data is aligned in the time dimension.

[0021] The raw sensor data is denoised, filtered, and feature extracted to obtain multi-source feature data, which includes material identification information, material size, and barcode features.

[0022] Timestamps and multi-source feature data are encapsulated in a unified JSON format to generate collaborative data packages, which are then synchronously uploaded to the central scheduling engine.

[0023] As a further aspect of the present invention: in the step of constructing a global state diagram of the real-time status of the sorting system through a central intelligent scheduling engine, and updating the state vector and attribute vector of each material in the global state diagram based on a collaborative data table:

[0024] The global state graph is represented by a directed graph G=(V,E), where V is a set of vertices used to represent the conveyor belt inlet node, conveyor belt intersection node, sorting port node and conveyor belt outlet node, and E is a set of edges used to represent the feasible conveyor paths between nodes. Each edge is accompanied by an attribute vector, which includes the path length, the current conveyor belt speed and the path occupancy status.

[0025] The global state diagram is updated synchronously at a fixed period.

[0026] As a further aspect of the present invention: the step of performing path deduction on in-transit materials based on the updated global state diagram, determining whether a path conflict has occurred, and calculating the corresponding conflict severity index specifically includes:

[0027] Based on the current global location coordinates of the material in transit, the speed of the conveyor belt, and the target sorting port, the position sequence within a future preset time window is simulated along the shortest path in the global state diagram.

[0028] Compare the location sequence of any pair of materials in transit. If the comparison result meets one of the preset conditions, it is determined to be a path conflict.

[0029] For materials in transit that are determined to have path conflicts, a conflict severity index is calculated based on the arrival time difference, load rate deviation, and material urgency. The conflict severity index is positively correlated with the processing priority.

[0030] As a further embodiment of the present invention: in the step of prioritizing rescheduling adjustment based on the conflict severity index and the load distribution of each sorting station, the priority rescheduling adjustment includes priority rearrangement strategy, path replanning strategy and speed control strategy, and also includes a material scheduling priority calculation step.

[0031] The material scheduling priority p_dynamic is characterized as follows:

[0032] p_dynamic=α×p_base+β×(1-1 / (1+S_conflict))+γ×(1-load_rate_target)+δ×urgency_factor,

[0033] Where p_base is the base priority, S_conflict is the highest conflict severity currently faced by the material, load_rate_target is the current load rate of the target sorting port, urgency_factor is the urgency factor, and α, β, γ, and δ are weighting coefficients.

[0034] This invention aims to provide an intelligent sorting and scheduling system based on multi-node information collaboration, comprising:

[0035] The data synchronization module is used to synchronize sensor data of multi-source information nodes through the edge collaboration gateway, and to perform edge preprocessing on the synchronized multi-source sensor data to obtain standardized collaborative data packets.

[0036] The status update module is used to construct a global status diagram of the real-time status of the sorting system through the central intelligent scheduling engine, and update the status vector and attribute vector of each material in the global status diagram based on the collaborative data table.

[0037] The conflict determination module is used to perform path deduction on materials in transit based on the updated global state diagram, determine whether a path conflict has occurred, and calculate the corresponding conflict severity index.

[0038] The optimization and adjustment module is used to perform priority rescheduling adjustments based on the conflict severity index and the load distribution of each sorting station in order to obtain and establish the optimal scheduling scheme for load balancing.

[0039] The scheduling and distribution module is used to convert the optimal scheduling scheme into scheduling instructions and send them to the execution device through the edge collaboration gateway to control the execution device to perform sorting actions;

[0040] The feedback update module is used to feed back the execution results through the edge collaboration gateway and update the global state graph based on the execution results.

[0041] As a further aspect of the present invention: the data synchronization module includes:

[0042] The data acquisition unit is used to synchronize sensor data from multi-source sensor nodes through an edge collaboration gateway. The sensor nodes include RFID tag reading nodes, visual inspection nodes, conveyor belt position sensor nodes, and sorting port load status sensor nodes.

[0043] The node synchronization unit is used to synchronize the timestamps of several sensor nodes associated with the edge collaborative gateway based on the clock reference and time synchronization protocol of the central intelligent scheduling engine, and to align the sensor data in the time dimension.

[0044] The data processing unit is used to denoise, filter, and extract features from the raw data of the sensor to obtain multi-source feature data, which includes material identification information, material size, and barcode features.

[0045] The data encapsulation unit is used to encapsulate timestamps and multi-source feature data in a unified JSON format, generate collaborative data packets, and synchronously upload them to the central scheduling engine.

[0046] As a further embodiment of the present invention: in the state update module:

[0047] The global state graph is represented by a directed graph G=(V,E), where V is a set of vertices used to represent the conveyor belt inlet node, conveyor belt intersection node, sorting port node and conveyor belt outlet node, and E is a set of edges used to represent the feasible conveyor paths between nodes. Each edge is accompanied by an attribute vector, which includes the path length, the current conveyor belt speed and the path occupancy status.

[0048] The global state diagram is updated synchronously at a fixed period.

[0049] As a further embodiment of the present invention: the conflict determination module includes:

[0050] The position simulation unit is used to simulate the position sequence within a future preset time window based on the current global position coordinates of the material in transit, the speed of the conveyor belt, and the target sorting port, along the shortest path in the global state diagram.

[0051] The location determination unit is used to compare and determine the location sequence of any pair of materials in transit. If the comparison and determination result meets one of the preset conditions, it is determined to be a path conflict.

[0052] The conflict calculation unit is used to calculate the conflict severity index for materials in transit that are determined to be in path conflicts, based on the arrival time difference, load rate deviation and the urgency of the materials. The conflict severity index is positively correlated with the processing priority.

[0053] As a further embodiment of the present invention: the priority rescheduling adjustment includes a priority rearrangement strategy, a path replanning strategy, and a speed control strategy, and the optimization adjustment module includes material scheduling priority;

[0054] The material scheduling priority p_dynamic is characterized as follows:

[0055] p_dynamic=α×p_base+β×(1-1 / (1+S_conflict))+γ×(1-load_rate_target)+δ×urgency_factor,

[0056] Where p_base is the base priority, S_conflict is the highest conflict severity currently faced by the material, load_rate_target is the current load rate of the target sorting port, urgency_factor is the urgency factor, and α, β, γ, and δ are weighting coefficients.

[0057] Compared with the prior art, the beneficial effects of the present invention include:

[0058] By eliminating information silos and achieving multi-node collaboration, this invention constructs an information aggregation architecture consisting of a multi-source information acquisition layer and an edge collaboration gateway layer. Combined with a precise time synchronization protocol and a unified spatial coordinate system, it integrates heterogeneous data scattered across RFID, vision, location, and load sensors into a unified material state vector. Compared to existing manual scheduling and single data source solutions, this invention achieves precise collaboration of multi-node information in the three dimensions of time, space, and semantics, fundamentally eliminating the problem of information silos.

[0059] Intelligent prediction and avoidance of sorting path conflicts: By using a path conflict prediction module to predict the material movement trajectory within the next 5 seconds, combined with conflict severity assessment and dynamic priority scheduling strategies (priority rearrangement, path replanning, and speed control), this invention achieves advance prediction and proactive avoidance of sorting path conflicts. Compared with the passive conflict handling methods in existing solutions, this invention can reduce the sorting path conflict rate by more than 80%, significantly reducing system downtime caused by conflicts.

[0060] To improve equipment resource utilization, the load balancing module monitors and dynamically adjusts the real-time load of each sorting station, ensuring that the load standard deviation is controlled within 15% to avoid some equipment being overloaded while others are idle. At the same time, the dynamic priority scheduling algorithm comprehensively considers the urgency of materials, the load of sorting stations, and the probability of path conflicts, thereby improving the overall equipment resource utilization rate by 30% to 50%.

[0061] Enhanced system adaptability, the closed-loop control architecture of perception-prediction-scheduling-execution-feedback enables real-time perception of changes in the sorting site's operating conditions (such as equipment failure, material surge, and sorting port congestion) and automatic adjustment of scheduling strategies. When a sorting port fails, the system automatically reroutes materials to a backup sorting port. When a certain type of material suddenly surges, the system automatically adjusts priority allocation and load balancing strategies to ensure the stability of overall sorting efficiency.

[0062] The entire process is automated, eliminating the need for manual intervention in vehicle identification, sorting, allocation, scheduling and coordination. Compared with the traditional manual sorting mode, it can reduce the number of operators by more than 70%, while reducing the sorting error rate from 3%-5% in the manual mode to less than 0.1%. Attached Figure Description

[0063] Figure 1 This is the overall process framework for an intelligent sorting and scheduling method based on multi-node information collaboration.

[0064] Figure 2 This is a system architecture diagram for an intelligent sorting and scheduling method based on multi-node information collaboration.

[0065] Figure 3 This is a flowchart of path conflict prediction and dynamic scheduling in an intelligent sorting and scheduling method based on multi-node information collaboration. Detailed Implementation

[0066] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0067] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0068] like Figure 2 As shown, this invention includes a multi-source information acquisition layer, an edge collaborative gateway layer, a central intelligent scheduling engine, and a sorting execution layer. The multi-source information acquisition layer includes RFID tag reading nodes, visual inspection nodes, conveyor belt position sensor nodes, and sorting port load status sensor nodes, responsible for acquiring material identity, physical attributes, real-time location, and sorting equipment status information from different physical locations and dimensions. The edge collaborative gateway layer is deployed at each sorting station, responsible for preprocessing, timestamp synchronization, and feature extraction of local multi-source sensor data, uploading standardized data to the central intelligent scheduling engine, and receiving scheduling instructions to issue to the execution equipment. The central intelligent scheduling engine is the core of this invention, maintaining a global material status graph, running a path conflict prediction model, and a dynamic priority scheduling algorithm to achieve globally optimal scheduling of sorting tasks. The sorting execution layer includes execution equipment such as conveyor belt sorting mechanisms, AGV carts, and robotic arms, receiving scheduling instructions to complete specific material sorting operations.

[0069] like Figure 1 , Figure 2 As shown, an intelligent sorting and scheduling method based on multi-node information collaboration is provided in one embodiment of the present invention, including the following steps:

[0070] S10, through the edge collaboration gateway, the sensor data of the multi-source information nodes are synchronized, and the synchronized multi-source sensor data is preprocessed at the edge to obtain standardized collaborative data packets;

[0071] S20, construct a global state diagram of the real-time status of the sorting system through the central intelligent scheduling engine, and update the state vector and attribute vector of each material in the global state diagram based on the collaborative data table;

[0072] S30, Based on the updated global state diagram, perform path deduction on the materials in transit, determine whether a path conflict has occurred, and calculate the corresponding conflict severity index;

[0073] S40, based on the conflict severity index and the load distribution of each sorting point, performs priority rescheduling adjustment to obtain and establish the optimal scheduling scheme for load balancing.

[0074] S50, the optimal scheduling scheme is converted into a scheduling instruction and sent to the execution device through the edge collaboration gateway to control the execution device to perform sorting actions;

[0075] S60, the execution result is fed back through the edge collaboration gateway, and the global state graph is updated based on the execution result.

[0076] This embodiment provides an intelligent sorting and scheduling method based on multi-node information collaboration. The core of this method lies in constructing a four-layer system architecture: a multi-source information acquisition layer, an edge collaboration gateway layer, a central intelligent scheduling engine layer, and a sorting execution layer. The edge collaboration gateway enables timestamp synchronization and feature alignment of multi-node data, while the central intelligent scheduling engine maintains a global state graph, implements a path conflict prediction model, and employs a dynamic priority scheduling algorithm. This achieves intelligent scheduling of sorting tasks and collaborative control of execution equipment. In this invention, the collaboration mechanism of multi-source information nodes achieves deep integration and collaborative utilization through three levels:

[0077] At the time coordination level: The edge coordination gateway adopts the IEEE 1588 PTP protocol and uses the GPS / BeiDou timing signal of the central scheduling engine as the primary clock source to synchronize and calibrate the local clock of each sensor node, ensuring that all collected data carries a unified and accurate timestamp, with a time synchronization accuracy of ±1 microsecond, which meets the real-time requirements of production logistics sorting and scheduling.

[0078] At the spatial collaboration level: a unified spatial coordinate system is established (with the starting end of the sorting line conveyor belt as the origin, the conveyor belt running direction as the X-axis, the vertical direction as the Y-axis, and the height direction as the Z-axis), and the physical installation positions of each sensor node are pre-calibrated in the unified coordinate system; during the data preprocessing stage, the edge collaboration gateway converts the location information collected by each sensor into the global spatial coordinate system in a unified manner, so as to achieve accurate alignment of multi-source data in spatial dimensions.

[0079] Semantic collaboration level: Define a unified data semantic model to map heterogeneous data (RFID identity data, visual feature data, location trajectory data, load status data) collected by different sensors to a unified material state vector S; the semantic mapping rules are pre-set in the configuration file of the edge collaboration gateway, supporting flexible expansion and dynamic updates.

[0080] In another preferred embodiment of the present invention, the step of synchronizing sensor data of multi-source information nodes through an edge collaboration gateway and performing edge preprocessing on the synchronized multi-source sensor data to obtain standardized collaborative data packets specifically includes:

[0081] Sensor data from multiple sensor nodes, including RFID tag reading nodes, visual inspection nodes, conveyor belt position sensor nodes, and sorting port load status sensor nodes, are synchronized through an edge collaboration gateway.

[0082] Based on the clock reference and time synchronization protocol of the central intelligent scheduling engine, the timestamps of several sensor nodes associated with the edge collaborative gateway are synchronized, and the sensor data is aligned in the time dimension.

[0083] The raw sensor data is denoised, filtered, and feature extracted to obtain multi-source feature data, which includes material identification information, material size, and barcode features.

[0084] Timestamps and multi-source feature data are encapsulated in a unified JSON format to generate collaborative data packages, which are then synchronously uploaded to the central scheduling engine.

[0085] In this embodiment, RFID tag reading nodes are deployed at the sorting line entrance and the intersection of each conveyor belt. They employ ultra-high frequency RFID readers (working frequency band 920-925MHz, adjustable reading distance 0-8m) to read RFID electronic tags attached to materials or bins, obtaining basic information such as the material's unique identifier (material ID), target sorting port number, material type, and urgency level. Visual inspection nodes are deployed at the sorting line entrance and key identification stations, using high-definition industrial cameras (resolution no less than 1920×1080, frame rate ≥30fps) combined with a deep learning object detection algorithm (based on an improved YOLOv8 model) to detect the material's external dimensions (length, width, height). Real-time detection and recognition of surface barcodes / QR codes and shape features (shape, color); incremental rotary encoders (resolution not less than 1000 pulses / revolution) are evenly deployed along the conveyor belt to collect the conveyor belt speed and the precise position information of the material on the conveyor belt in real time. The position information is encoded with the starting end of the conveyor belt as the origin and the precision unit is millimeters; the sorting port load status sensor nodes are deployed at each sorting port, using a combination of through-beam photoelectric sensors and weighing sensors to detect the current quantity of materials to be sorted, the stacking status and the load rate of each sorting port in real time. The load rate is calculated as: load rate = current quantity of materials to be sorted / maximum buffer capacity of sorting port × 100%.

[0086] The edge collaboration gateway layer is deployed on edge computing devices at each sorting station, responsible for aggregating, preprocessing, and synchronizing multi-source sensor data within the coverage area of ​​each station, forming standardized collaborative data packets that are then uploaded to the central intelligent scheduling engine. Multi-source data aggregation utilizes various industrial communication interfaces such as Ethernet, RS485, and CAN bus to gather raw data streams from RFID readers, industrial cameras, encoders, and photoelectric sensors. Timestamp synchronization and alignment adopt IEEE... The 1588 Precision Time Protocol (PTP) uses the clock of the central scheduling engine as a reference to synchronize data from each sensor node with nanosecond-level timestamps. Each sensor data point is attached with a unified timestamp tag to ensure precise alignment of multi-source data in the time dimension. Data preprocessing and feature extraction perform denoising, filtering, and feature extraction on the raw data. Visual data is denoised by Gaussian filtering (kernel size 3×3, σ=0.8), and the improved YOLOv8 model is used to extract material size and barcode features. RFID data is processed by CRC check and deduplication to extract material identity and target information. Encoder data is processed by Kalman filtering to output the precise position and velocity estimate of the material. Standardized data encapsulation encapsulates the processed multi-source data in a unified JSON format, including fields such as timestamp, material ID, current position coordinates (x,y,z), target sorting port, material size, urgency level, and data source type, and uploads it to the central scheduling engine via industrial Ethernet (Modbus TCP or OPC UA protocol).

[0087] In another preferred embodiment of the present invention, the step of constructing a global state diagram of the real-time status of the sorting system through a central intelligent scheduling engine, and updating the state vector and attribute vector of each material in the global state diagram based on a collaborative data table:

[0088] The global state graph is represented by a directed graph G=(V,E), where V is a set of vertices used to represent the conveyor belt inlet node, conveyor belt intersection node, sorting port node and conveyor belt outlet node, and E is a set of edges used to represent the feasible conveyor paths between nodes. Each edge is accompanied by an attribute vector, which includes the path length, the current conveyor belt speed and the path occupancy status.

[0089] The global state diagram is updated synchronously at a fixed period.

[0090] In this embodiment, each material is represented as a moving entity in the global state graph, with an attached state vector S=[material ID, current position coordinates, current edge e_current, target sorting port node v_target, estimated arrival time array for each node, current priority p_current, material size]. The global state graph is updated at a fixed period Δt=100ms. Each update receives standardized data packets from each edge collaborative gateway to update the state vector of the corresponding material and the attribute vector of the corresponding edge, ensuring the consistency between the global state graph and the actual physical state.

[0091] like Figure 3 As shown, in another preferred embodiment of the present invention, the step of performing path deduction on materials in transit based on the updated global state diagram, determining whether a path conflict has occurred, and calculating the corresponding conflict severity index specifically includes:

[0092] Based on the current global location coordinates of the material in transit, the speed of the conveyor belt, and the target sorting port, the position sequence within a future preset time window is simulated along the shortest path in the global state diagram.

[0093] Compare the location sequence of any pair of materials in transit. If the comparison result meets one of the preset conditions, it is determined to be a path conflict.

[0094] For materials in transit that are determined to have path conflicts, a conflict severity index is calculated based on the arrival time difference, load rate deviation, and material urgency. The conflict severity index is positively correlated with the processing priority.

[0095] In this embodiment, during the simulation, the future preset time window is within T_predict = 5 seconds. Specifically, the path prediction includes the following steps: For each material m_i in transit, based on its current position, the speed v_i of the conveyor belt, and the target sorting port v_target, a forward deduction is performed along the shortest path of the global state graph to calculate the expected position sequence P_i = [(t_1,pos_1),(t_2,pos_2),...,(t_k,pos_k)] of material m_i at each time point within the next T_predict seconds; the preset time window is determined. The conditions include: First, the two materials are expected to arrive at the same sorting port v_target at the same time t_c, and the arrival time difference Δt_arrival < safety interval Δt_safe (the default safety interval is 500ms). Second, the two materials are expected to meet at the conveyor belt intersection node at the same time t_c, that is, the time difference between the arrival of the two materials at the same intersection node is less than the safety interval. Third, when material m_i is expected to arrive at the sorting port v_target, the current load rate of the sorting port has exceeded the load limit threshold η_max (default 80%), and cannot accommodate new materials.

[0096] The conflict severity index S_conflict is characterized as follows:

[0097] S_conflict=w_1×(1-Δt_arrival / Δt_safe)+w_2×(load_rate-η_safe)+w_3×p_i;

[0098] Where w_1, w_2, and w_3 are weighting coefficients (default w_1=0.5, w_2=0.3, w_3=0.2), p_i is the material urgency level (levels 1-5), η_safe is the safe load rate threshold (default 60%), and the higher the conflict severity S_conflict, the more priority the conflict needs to be addressed.

[0099] like Figure 3 As shown, in another preferred embodiment of the present invention, the step of priority rescheduling adjustment based on the conflict severity index and the load distribution of each sorting point includes priority reordering strategy, path replanning strategy and speed control strategy, and also includes a material scheduling priority calculation step.

[0100] The material scheduling priority p_dynamic is characterized as follows:

[0101] p_dynamic=α×p_base+β×(1-1 / (1+S_conflict))+γ×(1-load_rate_target)+δ×urgency_factor,

[0102] Where p_base is the base priority, S_conflict is the highest conflict severity currently faced by the material, load_rate_target is the current load rate of the target sorting port, urgency_factor is the urgency factor, and α, β, γ, and δ are weighting coefficients.

[0103] In this embodiment, p_base is the basic priority determined by the material type and urgency level, with a value of 1-5. The urgency level factor has a value range of 0-1, with higher values ​​for closer deadlines. α, β, γ, and δ are weighting coefficients, with default values ​​of α=0.3, β=0.3, γ=0.2, and δ=0.2.

[0104] Priority reordering: Compare the dynamic priorities p_dynamic of conflicting materials. The material with lower priority is assigned a waiting time of Δt_delay. The formula for calculating Δt_delay is: Δt_delay = Δt_safe + path adjustment time.

[0105] Path replanning: When conflicting materials have the same target sorting port and the load rate exceeds η_max, check if there is an alternative backup sorting port (i.e., an idle sorting port that can handle the same type of materials); if so, reroute the lower priority material to the backup sorting port and update the target sorting port of the material in the global state graph.

[0106] Speed ​​adjustment: When conflicts cannot be resolved by priority rearrangement or path replanning, the arrival time of materials is staggered by adjusting the conveyor belt speed v. The adjusted speed v_adjust = v_current × (1 + Δt_offset / t_expected), where Δt_offset is the time offset to be staggered and t_expected is the expected arrival time.

[0107] As another preferred embodiment of the present invention, the load balancing strategy is as follows:

[0108] The load rate of all sorting ports is collected periodically (period T_balance = 2 seconds), and the global average load rate avg_load and the load rate standard deviation σ_load are calculated.

[0109] When σ_load > σ_threshold (default threshold 15%), load balancing adjustments are triggered.

[0110] For newly entered materials, when assigning target sorting ports, the sorting port with the lowest load rate among the same type should be selected first.

[0111] For materials that have been assigned but have not yet arrived at the sorting port, if the load rate of the target sorting port exceeds η_max and there is a spare sorting port of the same type with a load rate lower than η_safe, then the materials will be rerouted to the spare sorting port.

[0112] In another preferred embodiment of the present invention, the sorting execution layer includes execution devices such as a conveyor belt sorting mechanism, an AGV autonomous mobile vehicle, and an industrial robotic arm. It receives scheduling instructions from the central intelligent scheduling engine via industrial Ethernet or fieldbus (such as EtherCAT) to complete specific material sorting operations. The conveyor belt sorting mechanism uses a swing wheel or push-arm type sorting device with a response time ≤200ms and a sorting accuracy ≤±5mm. The AGV uses laser SLAM navigation with a positioning accuracy ≤±10mm and a maximum load capacity of 500kg. The industrial robotic arm uses a six-axis articulated structure with a repeatability accuracy ≤±0.1mm. Each execution device communicates with the central scheduling engine through a unified sorting instruction protocol. The instruction format includes fields such as: material ID, current conveyor belt position, target sorting port, execution action (sorting / pause / rerouting), and execution time limit.

[0113] In a specific embodiment of the present invention, the implementation steps can be summarized as follows:

[0114] RFID readers and industrial cameras are deployed at the entrance of the sorting line, encoders are installed at the conveyor drive end, and photoelectric sensors are installed at each sorting port; RFID tags are affixed to the surface of the materials (finished product bins), and the material ID, product type and target sorting port number are written in the tags.

[0115] Materials enter the sorting line via conveyor belt. RFID readers read tag information at a rate of 1100 tags / second. Industrial cameras simultaneously capture material images and detect material size using the YOLOv8 model (to confirm bin specifications). Encoders track material position in real time. The three data streams are synchronized and fused at the edge collaborative gateway to generate standardized collaborative data packets, which are then uploaded to the central scheduling engine.

[0116] The central scheduling engine updates the global state graph, recording the location and status of each material; the path conflict prediction module detects that the estimated arrival time difference between material A (target sorting port 3) and material B (target sorting port 3) is 200ms, which is less than the safety interval of 500ms, and determines it as a potential conflict.

[0117] The dynamic priority scheduling module calculates the priority of material A as p_dynamic=4.2 (urgent order) and the priority of material B as p_dynamic=2.8 (regular order). According to the priority reordering strategy, material B is assigned a waiting time of 500ms.

[0118] The load balancing module detected that the load rate of sorting port 3 was 85% (exceeding η_max=80%), while the load rate of sorting port 5 (also an accessory sorting port) was only 35%. For material C that newly enters the system and is destined for sorting port 3, the load balancing module reallocates its target to sorting port 5.

[0119] After the scheduling command is issued, the swing wheel sorting mechanism accurately pushes the material into the corresponding sorting port at the specified time. After execution, the photoelectric sensor confirms that the material has arrived and feeds back the result to the scheduling engine, completing the closed loop.

[0120] Implementation results: The system ran continuously for 8 hours, processed 24,000 finished products, achieved a sorting accuracy rate of 99.7%, reduced the path conflict rate from 12% in the manual scheduling mode to 1.5%, and increased the overall equipment utilization rate (OEE) from 62% to 87%.

[0121] This invention also provides an intelligent sorting and scheduling system based on multi-node information collaboration, which includes:

[0122] The data synchronization module is used to synchronize sensor data of multi-source information nodes through the edge collaboration gateway, and to perform edge preprocessing on the synchronized multi-source sensor data to obtain standardized collaborative data packets.

[0123] The status update module is used to construct a global status diagram of the real-time status of the sorting system through the central intelligent scheduling engine, and update the status vector and attribute vector of each material in the global status diagram based on the collaborative data table.

[0124] The conflict determination module is used to perform path deduction on materials in transit based on the updated global state diagram, determine whether a path conflict has occurred, and calculate the corresponding conflict severity index.

[0125] The optimization and adjustment module is used to perform priority rescheduling adjustments based on the conflict severity index and the load distribution of each sorting station in order to obtain and establish the optimal scheduling scheme for load balancing.

[0126] The scheduling and distribution module is used to convert the optimal scheduling scheme into scheduling instructions and send them to the execution device through the edge collaboration gateway to control the execution device to perform sorting actions;

[0127] The feedback update module is used to feed back the execution results through the edge collaboration gateway and update the global state graph based on the execution results.

[0128] In another preferred embodiment of the present invention, the data synchronization module includes:

[0129] The data acquisition unit is used to synchronize sensor data from multi-source sensor nodes through an edge collaboration gateway. The sensor nodes include RFID tag reading nodes, visual inspection nodes, conveyor belt position sensor nodes, and sorting port load status sensor nodes.

[0130] The node synchronization unit is used to synchronize the timestamps of several sensor nodes associated with the edge collaborative gateway based on the clock reference and time synchronization protocol of the central intelligent scheduling engine, and to align the sensor data in the time dimension.

[0131] The data processing unit is used to denoise, filter, and extract features from the raw data of the sensor to obtain multi-source feature data, which includes material identification information, material size, and barcode features.

[0132] The data encapsulation unit is used to encapsulate timestamps and multi-source feature data in a unified JSON format, generate collaborative data packets, and synchronously upload them to the central scheduling engine.

[0133] In another preferred embodiment of the present invention, the state update module includes:

[0134] The global state graph is represented by a directed graph G=(V,E), where V is a set of vertices used to represent the conveyor belt inlet node, conveyor belt intersection node, sorting port node and conveyor belt outlet node, and E is a set of edges used to represent the feasible conveyor paths between nodes. Each edge is accompanied by an attribute vector, which includes the path length, the current conveyor belt speed and the path occupancy status.

[0135] The global state diagram is updated synchronously at a fixed period.

[0136] In another preferred embodiment of the present invention, the conflict determination module includes:

[0137] The position simulation unit is used to simulate the position sequence within a future preset time window based on the current global position coordinates of the material in transit, the speed of the conveyor belt, and the target sorting port, along the shortest path in the global state diagram.

[0138] The location determination unit is used to compare and determine the location sequence of any pair of materials in transit. If the comparison and determination result meets one of the preset conditions, it is determined to be a path conflict.

[0139] The conflict calculation unit is used to calculate the conflict severity index for materials in transit that are determined to be in path conflicts, based on the arrival time difference, load rate deviation and the urgency of the materials. The conflict severity index is positively correlated with the processing priority.

[0140] As another preferred embodiment of the present invention, the priority rescheduling adjustment includes a priority rearrangement strategy, a path replanning strategy, and a speed control strategy, and the optimization adjustment module includes material scheduling priority;

[0141] The material scheduling priority p_dynamic is characterized as follows:

[0142] p_dynamic=α×p_base+β×(1-1 / (1+S_conflict))+γ×(1-load_rate_target)+δ×urgency_factor,

[0143] Where p_base is the base priority, S_conflict is the highest conflict severity currently faced by the material, load_rate_target is the current load rate of the target sorting port, urgency_factor is the urgency factor, and α, β, γ, and δ are weighting coefficients.

[0144] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0145] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the disclosure in the specification and embodiments. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0146] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. An intelligent sorting and scheduling method based on multi-node information collaboration, characterized in that, Include: The sensor data of multi-source information nodes is synchronized through the edge collaboration gateway, and the synchronized multi-source sensor data is preprocessed at the edge to obtain standardized collaborative data packets. A global state diagram of the sorting system's real-time status is constructed through a central intelligent scheduling engine, and the state vectors and attribute vectors of each material in the global state diagram are updated based on a collaborative data table. Based on the updated global state diagram, the path of materials in transit is simulated to determine whether a path conflict has occurred and to calculate the corresponding conflict severity index. Priority rescheduling is performed based on the conflict severity index and the load distribution of each sorting station in order to obtain and establish the optimal load-balanced scheduling scheme. The optimal scheduling scheme is converted into scheduling instructions and sent to the execution device through the edge collaboration gateway to control the execution device to perform sorting actions; The execution results are fed back through the edge collaboration gateway, and the global state graph is updated based on the execution results.

2. The intelligent sorting and scheduling method based on multi-node information collaboration according to claim 1, characterized in that, The steps of synchronizing sensor data from multiple source information nodes through an edge collaboration gateway and performing edge preprocessing on the synchronized multi-source sensor data to obtain standardized collaborative data packets specifically include: Sensor data from multiple sensor nodes, including RFID tag reading nodes, visual inspection nodes, conveyor belt position sensor nodes, and sorting port load status sensor nodes, are synchronized through an edge collaboration gateway. Based on the clock reference and time synchronization protocol of the central intelligent scheduling engine, the timestamps of several sensor nodes associated with the edge collaborative gateway are synchronized, and the sensor data is aligned in the time dimension. The raw sensor data is denoised, filtered, and feature extracted to obtain multi-source feature data, which includes material identification information, material size, and barcode features. Timestamps and multi-source feature data are encapsulated in a unified JSON format to generate collaborative data packages, which are then synchronously uploaded to the central scheduling engine.

3. The intelligent sorting and scheduling method based on multi-node information collaboration according to claim 2, characterized in that, In the step of constructing a global state diagram of the real-time status of the sorting system through a central intelligent scheduling engine, and updating the state vector and attribute vector of each material in the global state diagram based on a collaborative data table: The global state graph is represented by a directed graph G=(V,E), where V is a set of vertices used to represent the conveyor belt inlet node, conveyor belt intersection node, sorting port node and conveyor belt outlet node, and E is a set of edges used to represent the feasible conveyor paths between nodes. Each edge is accompanied by an attribute vector, which includes the path length, the current conveyor belt speed and the path occupancy status. The global state diagram is updated synchronously at a fixed period.

4. The intelligent sorting and scheduling method based on multi-node information collaboration according to claim 3, characterized in that, The step of performing path deduction for materials in transit based on the updated global state diagram, determining whether path conflicts have occurred, and calculating the corresponding conflict severity index specifically includes: Based on the current global location coordinates of the material in transit, the speed of the conveyor belt, and the target sorting port, the position sequence within a future preset time window is simulated along the shortest path in the global state diagram. Compare the location sequence of any pair of materials in transit. If the comparison result meets one of the preset conditions, it is determined to be a path conflict. For materials in transit that are determined to have path conflicts, a conflict severity index is calculated based on the arrival time difference, load rate deviation, and material urgency. The conflict severity index is positively correlated with the processing priority.

5. The intelligent sorting and scheduling method based on multi-node information collaboration according to claim 4, characterized in that, The step of prioritizing rescheduling based on the conflict severity index and the load distribution of each sorting station includes priority rescheduling strategy, path replanning strategy and speed control strategy, as well as the calculation step of material scheduling priority. The material scheduling priority p_dynamic is characterized as follows: p_dynamic=α×p_base+β×(1-1 / (1+S_conflict))+γ×(1-load_rate_target)+δ×urgency_factor, Where p_base is the base priority, S_conflict is the highest conflict severity currently faced by the material, load_rate_target is the current load rate of the target sorting port, urgency_factor is the urgency factor, and α, β, γ, and δ are weighting coefficients.

6. An intelligent sorting and scheduling system based on multi-node information collaboration, characterized in that, Include: The data synchronization module is used to synchronize sensor data of multi-source information nodes through the edge collaboration gateway, and to perform edge preprocessing on the synchronized multi-source sensor data to obtain standardized collaborative data packets. The status update module is used to construct a global status diagram of the real-time status of the sorting system through the central intelligent scheduling engine, and update the status vector and attribute vector of each material in the global status diagram based on the collaborative data table. The conflict determination module is used to perform path deduction on materials in transit based on the updated global state diagram, determine whether a path conflict has occurred, and calculate the corresponding conflict severity index. The optimization and adjustment module is used to perform priority rescheduling adjustments based on the conflict severity index and the load distribution of each sorting station in order to obtain and establish the optimal scheduling scheme for load balancing. The scheduling and distribution module is used to convert the optimal scheduling scheme into scheduling instructions and send them to the execution device through the edge collaboration gateway to control the execution device to perform sorting actions; The feedback update module is used to feed back the execution results through the edge collaboration gateway and update the global state graph based on the execution results.

7. The intelligent sorting and scheduling system based on multi-node information collaboration according to claim 6, characterized in that, The data synchronization module includes: The data acquisition unit is used to synchronize sensor data from multi-source sensor nodes through an edge collaboration gateway. The sensor nodes include RFID tag reading nodes, visual inspection nodes, conveyor belt position sensor nodes, and sorting port load status sensor nodes. The node synchronization unit is used to synchronize the timestamps of several sensor nodes associated with the edge collaborative gateway based on the clock reference and time synchronization protocol of the central intelligent scheduling engine, and to align the sensor data in the time dimension. The data processing unit is used to denoise, filter, and extract features from the raw data of the sensor to obtain multi-source feature data, which includes material identification information, material size, and barcode features. The data encapsulation unit is used to encapsulate timestamps and multi-source feature data in a unified JSON format, generate collaborative data packets, and synchronously upload them to the central scheduling engine.

8. The intelligent sorting and scheduling system based on multi-node information collaboration according to claim 7, characterized in that, In the state update module: The global state graph is represented by a directed graph G=(V,E), where V is a set of vertices used to represent the conveyor belt inlet node, conveyor belt intersection node, sorting port node and conveyor belt outlet node, and E is a set of edges used to represent the feasible conveyor paths between nodes. Each edge is accompanied by an attribute vector, which includes the path length, the current conveyor belt speed and the path occupancy status. The global state diagram is updated synchronously at a fixed period.

9. The intelligent sorting and scheduling system based on multi-node information collaboration according to claim 8, characterized in that, The conflict determination module includes: The position simulation unit is used to simulate the position sequence within a future preset time window based on the current global position coordinates of the material in transit, the speed of the conveyor belt, and the target sorting port, along the shortest path in the global state diagram. The location determination unit is used to compare and determine the location sequence of any pair of materials in transit. If the comparison and determination result meets one of the preset conditions, it is determined to be a path conflict. The conflict calculation unit is used to calculate the conflict severity index for materials in transit that are determined to be in path conflicts, based on the arrival time difference, load rate deviation and the urgency of the materials. The conflict severity index is positively correlated with the processing priority.

10. The intelligent sorting and scheduling system based on multi-node information collaboration according to claim 9, characterized in that, The priority rescheduling adjustment includes priority rearrangement strategy, path replanning strategy and speed control strategy, and the optimization adjustment module includes material scheduling priority; The material scheduling priority p_dynamic is characterized as follows: p_dynamic=α×p_base+β×(1-1 / (1+S_conflict))+γ×(1-load_rate_target)+δ×urgency_factor, Where p_base is the base priority, S_conflict is the highest conflict severity currently faced by the material, load_rate_target is the current load rate of the target sorting port, urgency_factor is the urgency factor, and α, β, γ, and δ are weighting coefficients.

Images