Modular intelligent drive card-based integrated control system for logistics conveying line

Abstract

The application belongs to the technical field of logistics automation, and discloses a logistics conveying line integrated control system based on a modular intelligent drive card; the system comprises a modular intelligent drive card, a conveying line section, a main control unit and a business host. The system collects equipment object running parameters and state data through the modular intelligent drive card, generates a running characteristic vector, classifies running states by using a dynamic clustering algorithm, generates running trend prediction data through a prediction model in combination with high-level business logic of the business host, generates and issues optimized control instructions based on the prediction data, generates a dynamic handover strategy through a network communication module to coordinate product handover between adjacent conveying line sections, and dynamically binds product tracking data and physical product positions through a virtual mapping mechanism. The application realizes efficient integrated control of the logistics conveying line, improves system response speed and processing capacity, optimizes energy utilization, and enhances product tracking accuracy.

Description

Technical Field

[0001] This invention relates to the field of logistics automation technology, and more specifically, to an integrated control system for logistics conveyor lines based on modular intelligent drive cards. Background Technology

[0002] With the rapid development of e-commerce and intelligent manufacturing, modern logistics systems are facing unprecedented challenges and opportunities. As a core component of the logistics system, the advancement of control technology in logistics conveyor lines directly impacts the efficiency and reliability of the entire system. In recent years, logistics conveyor line control systems have evolved from traditional mechanical control to electrical automation control, and then to information-based intelligent control.

[0003] Traditional logistics conveyor control systems primarily employ a PLC (Programmable Logic Controller)-based control architecture, connecting control units in various sections via fieldbus to achieve basic conveying functions. While this control method performs adequately in simple logistics scenarios, its limitations become increasingly apparent as logistics scales up and processing demands become more complex.

[0004] Traditional systems typically employ a distributed architecture, with each control unit operating independently and lacking a unified integration framework. This results in poor coordination between system components and an inability to achieve global optimization control. In high-speed logistics distribution centers, a large number of items need to be sorted and delivered within a short timeframe. However, existing systems suffer from lengthy data transmission links and significant delays in control command issuance, failing to meet millisecond-level response requirements and causing item backlogs and delivery delays. The product handover process between adjacent conveyor line sections lacks an intelligent scheduling mechanism, relying solely on simple fixed priority strategies. This fails to address dynamically changing handover demands, frequently leading to congestion at handover points and product collision damage. Furthermore, traditional systems use fixed power allocation methods, unable to dynamically adjust power supply strategies based on conveyor line load. This results in power shortages during peak periods and energy waste during off-peak periods, significantly increasing operating costs. Finally, the system lacks predictive analysis capabilities for equipment operating trends, only passively responding to existing faults without providing early warnings or preventative measures, leading to frequent unplanned equipment downtime. In terms of product tracking, traditional systems use discrete detection point tracking methods, which cannot achieve continuous position monitoring of items on the conveyor line. When the logistics density increases, problems such as item identification confusion and tracking loss can easily occur, seriously affecting the accuracy and reliability of logistics management. In addition, the systems generally lack environmental adaptability and cannot automatically adjust control parameters according to factors such as changes in temperature and humidity and equipment aging. They are unstable in different operating environments and have high maintenance costs.

[0005] In view of this, the present invention proposes an integrated control system for logistics conveyor lines based on modular intelligent drive cards to solve the above problems. Summary of the Invention

[0006] To overcome the aforementioned deficiencies of the prior art and to achieve the above objectives, the present invention provides the following technical solution: an integrated control system for a logistics conveyor line based on a modular intelligent drive card, comprising:

[0007] At least one modular intelligent drive card, at least one conveyor line section, at least one main control unit, and at least one service host;

[0008] The modular smart drive card includes a power management module and a network communication module. The power management module is used to provide power, perform dynamic power allocation and collect operating status data. The network communication module is used to realize multi-protocol data communication and real-time handover control.

[0009] The conveyor section includes at least one device object, which includes a drive motor, a sensor, and an actuator.

[0010] The main control unit is used to communicate with the modular intelligent drive card and to control the conveyor line section in real time through a standardized communication interface;

[0011] The business host is used to execute high-level business logic, including logistics task scheduling, data storage, and product tracking;

[0012] The system achieves integrated control of the logistics conveyor line through the following steps:

[0013] The operating parameters and status data of each device in the conveyor line section are obtained, and the corresponding operating feature vector is generated through the modular intelligent drive card.

[0014] Based on the operating feature vector, the operating status of the conveyor line section is classified by a dynamic clustering algorithm to generate a set of classified operating statuses.

[0015] Based on the set of operating states and the high-level business logic of the business host, predictive data of the operating trend of the conveyor line section is generated by a predictive model.

[0016] Based on the predicted operating trend data, optimization control commands are generated and issued through the standardized communication interface to adjust the operating parameters of the device object.

[0017] Based on the product handover requirements between the conveyor line sections, a dynamic handover strategy is generated through the network communication module, and the product handover between adjacent conveyor line sections is coordinated based on the dynamic handover strategy.

[0018] Based on the high-level business logic of the business host, product tracking data is generated, and the product tracking data is dynamically bound to the physical product location in the conveyor line section through a virtual mapping mechanism.

[0019] The technical effects and advantages of the integrated control system for logistics conveyor lines based on modular intelligent drive cards in this invention are as follows:

[0020] This invention improves logistics processing efficiency, making sorting and distribution processes smoother and more efficient, effectively alleviating logistics pressure during peak periods. In practical applications, the speed of goods transportation is significantly increased, waiting time is noticeably shortened, and logistics congestion is effectively controlled. The system optimizes the handover process between adjacent sections, making the handover process smoother, reducing damage to goods and delivery delays caused by poor handover, and improving the safety and reliability of the logistics system. In terms of energy utilization, the system achieves optimal resource allocation, significantly reducing energy consumption while ensuring normal operation, creating considerable economic benefits for enterprises. Through advanced fault prediction capabilities, the system significantly reduces unplanned equipment downtime, extends equipment lifespan, and reduces maintenance costs. High-precision item tracking ensures that each item can be accurately located during transportation, effectively eliminating the risk of lost or misdelivered items and improving customer satisfaction. The system's adaptive characteristics enable it to maintain stable and efficient operation under different environmental conditions without frequent manual intervention, reducing the workload of maintenance personnel. Furthermore, the system's flexible scalability allows logistics companies to easily upgrade according to business development needs, avoiding the huge costs and risks of a complete system replacement. Attached Figure Description

[0021] Figure 1 This is a schematic diagram of the integrated control system for a logistics conveyor line based on a modular intelligent drive card according to the present invention.

[0022] Figure 2 This is a schematic diagram illustrating the integrated control logic for the logistics conveyor line implemented in this invention. Detailed Implementation

[0023] 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.

[0024] This application provides an integrated control system for a logistics conveyor line based on modular intelligent drive cards. The execution entities of this integrated control system include, but are not limited to, systems such as: a logistics control center, a conveyor line management system, an intelligent warehousing system, an equipment monitoring platform, and an operation optimization system, which can be considered as general computing nodes in this application. The data processing platform includes, but is not limited to, at least one of: an equipment status monitoring system, an operation characteristic analysis system, and a handover control system.

[0025] Please see Figure 1 This invention provides an integrated control system for a logistics conveyor line based on a modular intelligent drive card, comprising at least one modular intelligent drive card, at least one conveyor line section, at least one main control unit, and at least one business host. The modular intelligent drive card includes a power management module and a network communication module. The power management module provides power, performs dynamic power allocation, and collects operating status data. The network communication module enables multi-protocol data communication and real-time handover control. The conveyor line section includes at least one device object, which includes a drive motor, sensors, and actuators. The main control unit communicates with the modular intelligent drive card and performs real-time control of the conveyor line section through a standardized communication interface. The business host executes high-level business logic, including logistics task scheduling, data storage, and product tracking.

[0026] This invention achieves intelligent control of logistics conveyor lines through modular design, constructs equipment object operation feature vectors to provide a data foundation for subsequent analysis, and performs dynamic clustering and state classification based on the operation feature vectors to make the control adaptive. The dynamic clustering algorithm can identify changes in operating modes under different working conditions, and trend prediction combined with the high-level business logic of the business host improves the foresight of the control. Optimizing the generation and issuance of control commands ensures the real-time response capability of the system. The network communication module realizes dynamic handover between adjacent conveyor line sections, solving the coordination problem of handover links in traditional conveyor systems. The dynamic binding of product tracking data with physical product location ensures the visualization and traceability of the logistics process, and the virtual mapping mechanism enhances the accuracy and real-time performance of product positioning.

[0027] Please see Figure 2 This is a schematic diagram illustrating the integrated control logic of the logistics conveyor line implemented in this application. In this embodiment of the invention, the system achieves integrated control of the logistics conveyor line through the following steps:

[0028] Step 1: Obtain the operating parameters and status data of each device in the conveyor line section, and generate the corresponding operating feature vector through the modular intelligent drive card;

[0029] Step 2: Based on the running feature vector, classify the running status of the transport line sections using a dynamic clustering algorithm to generate a set of classified running statuses;

[0030] Step 3: Based on the set of operating statuses and the high-level business logic of the business host, generate predictive data of the operating trend of the conveyor line section through a predictive model;

[0031] Step 4: Based on the predicted operational trends, generate and issue optimization control commands through a standardized communication interface to adjust the operating parameters of the equipment.

[0032] Step 5: Based on the product handover requirements between conveyor line sections, generate a dynamic handover strategy through the network communication module, and coordinate the product handover between adjacent conveyor line sections based on the dynamic handover strategy.

[0033] Step 6: Generate product tracking data based on the high-level business logic of the business host, and dynamically bind the product tracking data to the physical product location in the conveyor line section through a virtual mapping mechanism.

[0034] In this embodiment, the operating parameters of each device in the conveyor line section are first obtained, including parameters such as the speed, current, and temperature of the drive motor, data such as the position detection value, pressure value, and photoelectric status of the sensor, and information such as the action status, response time, and torque output of the actuator, forming a set of device operating parameters. Simultaneously, the status data of the device is collected, including working status (normal / abnormal / fault), cumulative running time, maintenance records, etc., forming a device status dataset. The collected operating parameters and status data are preliminarily processed by the data processing unit built into the modular intelligent drive card, including data normalization, outlier filtering, and signal smoothing, to ensure data quality. The processed data is then subjected to feature extraction using predefined feature extraction algorithms (such as time-domain statistical features, frequency-domain transform features, and time-frequency joint analysis features) to generate feature vectors that characterize the device's operating status. These feature vectors contain multiple dimensions, such as stability indicators, energy efficiency indicators, responsiveness indicators, and synchronization indicators, comprehensively depicting the device's operating characteristics and forming an operating feature vector.

[0035] Then, based on the operational feature vectors, a dynamic clustering algorithm is used to classify the operational status of the conveyor line sections. Unlike traditional static clustering methods, dynamic clustering can adaptively adjust cluster centers and category boundaries to adapt to the dynamic changes in the conveyor line's operational status. The algorithm first initializes cluster centers, which can be set based on historical data or expert experience, such as typical categories like normal operation, minor anomalies, and severe anomalies. For each time window, the similarity (e.g., Euclidean distance, Mahalanobis distance, cosine similarity) between the operational feature vector and each cluster center is calculated, and the feature vector is assigned to the most similar cluster. As new data is continuously generated, the cluster center positions and the number of clusters are dynamically adjusted so that the clustering results reflect the real-time changes in the conveyor line's operational status. The algorithm employs an adaptive threshold mechanism: when the variability of samples within a category exceeds a threshold, splitting to form a new category is considered; when the distance between the centers of two categories is less than a threshold, merging the categories is considered. This dynamic adjustment mechanism enables precise classification of the conveyor line's operating status, forming a set of classified operating statuses, such as normal operation, low energy efficiency, early signs of failure, and equipment aging.

[0036] Next, based on the categorized set of operating states and the high-level business logic provided by the business host (such as logistics task planning, product priority, capacity requirements, etc.), predictive models generate operational trend forecast data for the conveyor line sections. The predictive models employ deep learning architectures, such as Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU), or Temporal Convolutional Networks (TCN), to capture the temporal dependencies of the conveyor line's operating data. During model training, historical operating data and current business requirements are considered simultaneously to model multiple prediction objectives, such as equipment status trend prediction (predicting when equipment might fail), efficiency trend prediction (predicting future changes in conveyor efficiency), and energy consumption trend prediction (predicting fluctuations in energy consumption). The prediction results are presented in time-series format, including predicted values, confidence intervals, and multiple possible development paths, forming operational trend forecast data that provides a scientific basis for subsequent control optimization.

[0037] Based on operational trend prediction data, optimized control commands are generated and issued through a standardized communication interface. The generation process employs advanced control methods such as Model Predictive Control (MPC) or Reinforcement Learning (RL) to optimize control objectives (e.g., maximizing energy efficiency, minimizing product damage, and optimizing handover time) while meeting system constraints. These control commands are then sent to equipment in each conveyor line section via the standardized communication interface, enabling precise adjustments to operating parameters such as drive motor speed, actuator timing, and sensor sensitivity. The execution effect of the control commands is monitored in real-time through a feedback mechanism, forming a closed-loop control system to ensure stable and efficient system operation.

[0038] Based on the product handover requirements between conveyor line sections (such as product arrival time, product characteristics, handover priority, etc.), a dynamic handover strategy is generated through the network communication module. The handover strategy generation process considers multiple factors, such as the output capacity of the upstream conveyor line, the receiving capacity of the downstream conveyor line, product backlog, and physical limitations of the handover point. Using game theory or multi-agent collaborative decision-making methods, optimal handover parameters, such as handover time window, handover speed matching, and product interval control, are calculated to form the dynamic handover strategy. Based on the generated handover strategy, the product handover process between adjacent conveyor line sections is coordinated to achieve seamless handover and smooth transmission, avoiding problems such as congestion, collisions, or product damage.

[0039] Finally, based on the high-level business logic of the business host (such as order information, product attributes, delivery requirements, etc.), product tracking data is generated, including information such as the product's unique identifier, current location, historical trajectory, and estimated arrival time. Through a virtual mapping mechanism, the product tracking data is dynamically bound to the physical product location within the conveyor line segment. This virtual mapping mechanism, based on sensor networks and location estimation algorithms, updates the precise location of the product on the conveyor line in real time and keeps it synchronized with the tracking database, ensuring the accuracy and real-time nature of the tracking data. This binding mechanism enables visualized monitoring and precise tracking of products throughout the entire process, improving the transparency and management efficiency of the logistics system.

[0040] In this embodiment of the invention, the standardized communication interface includes a vertical interface and a horizontal interface, wherein:

[0041] The vertical interface includes a control interface and a reporting interface. The control interface is used to send optimized control commands from the main control unit to the modular intelligent drive card, and the reporting interface is used to upload operating status data and operating feature vectors from the modular intelligent drive card to the main control unit.

[0042] The horizontal interface includes a dynamic handover interface, which is used to define the product handover protocol between adjacent conveyor line sections. The product handover protocol includes dynamic request signals, dynamic response signals, handover priority signals, and extended auxiliary signals.

[0043] The dynamic handover interface uses a variable-length data frame structure, which carries product handover information, product tracking data, and priority information for dynamic handover strategies.

[0044] In this embodiment, the standardized communication interface adopts a layered design, dividing communication functions into two main categories: vertical interfaces and horizontal interfaces. The vertical interface is responsible for data exchange between upper and lower levels, including two sub-interfaces: a control interface and a reporting interface. The control interface adopts a command-response model, defining a standardized instruction set, including basic control instructions (such as start, stop, speed adjustment), parameter setting instructions (such as modifying operating parameters, updating thresholds), and mode switching instructions (such as switching to manual mode, diagnostic mode), etc. The instruction format is unified, including fields such as instruction header, target address, instruction code, parameter fields, and checksum, ensuring the reliability and accuracy of instruction transmission. The main control unit issues optimized control instructions to the modular intelligent drive card through the control interface to achieve precise control of the conveyor line equipment. The reporting interface adopts a publish-subscribe model, with the modular intelligent drive card periodically or event-triggered uploading operating status data and operating feature vectors to the main control unit. Different priorities and frequencies can be set for data uploads, such as high-priority real-time uploads of critical status data and low-priority batch uploads of ordinary monitoring data. The report data is in a structured format, including fields such as timestamp, device ID, data type, data value, and quality label, which facilitates data processing and analysis by the main control unit.

[0045] The horizontal interface is primarily responsible for collaborative communication between peer devices, with the dynamic handover interface as its core component. The dynamic handover interface defines the product handover protocol between adjacent conveyor line sections, enabling seamless collaboration between devices. The product handover protocol includes four key signals: a dynamic request signal, sent by the upstream device, containing information about the product to be handed over (e.g., size, weight, arrival time) and desired handover parameters (e.g., speed, interval); a dynamic response signal, replied by the downstream device, containing reception capability status (e.g., receiveable, delayed reception, rejected reception) and suggested handover parameters; a handover priority signal, used to determine the processing order when multiple handover requests conflict, with priority dynamically calculated based on product attributes (e.g., urgency, value) and system status; and extended auxiliary signals, used to transmit additional handover-related information, such as special processing requirements, product tracking data, and quality inspection results. These signals together constitute a complete handover protocol framework, supporting flexible and varied handover scenarios.

[0046] The dynamic handover interface uses a variable-length data frame structure, ensuring the complete transmission of necessary information while avoiding bandwidth waste caused by fixed-length frames. The data frame consists of a frame header (containing synchronization words, frame length, frame type, etc.), a frame body (containing handover information, tracking data, priority information, etc.), and a frame trailer (containing checksum, end marker, etc.). The frame body uses TLV (Type-Length-Value) encoding, allowing for flexible addition or deletion of information fields to adapt to different handover scenarios. Priority information is encoded in specific fields to guide the priority order of network transmission and receiver processing, ensuring that important handover information is not delayed due to network congestion. The entire data frame structure design balances real-time performance, reliability, and scalability, providing a reliable communication foundation for product handover.

[0047] In this embodiment of the invention, the method for generating feature vectors includes:

[0048] Obtain the operating parameters of each device in the conveyor line section, including motor speed, sensor readings, and actuator status;

[0049] The operating parameters are normalized to generate a set of normalized operating parameters;

[0050] Based on the normalized set of operating parameters, the operating fluctuation characteristics of the equipment object within a preset time window are extracted. The operating fluctuation characteristics include the variance, frequency distribution, and peak change rate of the operating parameters.

[0051] By fusing operational fluctuation characteristics with status data, an operational feature vector of the device object is generated.

[0052] In this embodiment, detailed operating parameters are first collected from each piece of equipment in the conveyor line section. For drive motors, parameters such as rotational speed (RPM), phase current (A), phase voltage (V), output torque (N·m), coil temperature (°C), and vibration intensity (mm / s) are collected. For sensors, readings such as coordinate values ​​(mm) of position sensors, trigger status (on / off) of photoelectric sensors, pressure values ​​(Pa) of pressure sensors, temperature values ​​(°C) of temperature sensors, and speed values ​​(m / s) of speed sensors are collected. For actuators, status information such as cylinder extension / retraction position (mm), solenoid valve on / off status (on / off), servo motor angle value (°), and sorting mechanism operation status (working / standby) are collected. These parameters are acquired in real time through the data acquisition interface of the modular intelligent drive card. The sampling frequency is dynamically adjusted according to the importance and rate of change of the parameters, such as high-frequency sampling (e.g., 100Hz) for key parameters and low-frequency sampling (e.g., 10Hz) for general parameters.

[0053] The acquired operating parameters are normalized to eliminate dimensional and numerical range differences between parameters, making subsequent analysis more objective and accurate. Normalization methods include: min-max normalization, which maps parameter values ​​to the [0,1] interval; converting parameters into a distribution with a mean of 0 and a standard deviation of 1; and logarithmic transformation, which compresses the range for parameters with large variations. Different parameters may require different normalization methods; the method best suited to the parameter's distribution characteristics is selected. All normalized parameters constitute a normalized operating parameter set, providing a standardized data foundation for subsequent feature extraction.

[0054] Based on a normalized set of operating parameters, the operational fluctuation characteristics of the equipment are extracted within a preset time window (e.g., 5 minutes, 30 minutes, or longer). These fluctuation characteristics reflect the changing patterns of the parameters in the time domain, including: variance characteristics, which calculate the statistical variance of the parameters within the window, reflecting the severity of parameter fluctuations; frequency distribution characteristics, which extract spectral features such as dominant frequency components and energy distribution by converting the time-domain signal to the frequency domain using Fast Fourier Transform (FFT) or wavelet transform; and peak value change rate, which calculates the frequency of parameter peaks and the rate of change between adjacent peaks, reflecting the stability of the system response. These fluctuation characteristics can capture the dynamic characteristics of equipment operation and are highly sensitive to abnormal states.

[0055] Finally, the extracted operational fluctuation features are fused with the equipment's status data (such as operating modes, fault records, maintenance cycles, etc.), and a comprehensive operational feature vector is generated using feature fusion algorithms (such as feature-level fusion or decision-level fusion). The feature vector dimensions are dynamically adjusted according to application requirements and computing resources, typically containing 10-50 feature dimensions, each representing a specific aspect of the equipment's operational status. To improve the expressive power of the feature vector, dimensionality reduction techniques (such as Principal Component Analysis (PCA) and Linear Discriminant Analysis (LDA) can be used to remove redundant information, or feature enhancement techniques (such as constructing combined features) can be used to enhance key information. The final generated operational feature vector is a high-dimensional abstract expression of the equipment's operational status, capable of comprehensively and accurately depicting the equipment's health status and performance level, providing strong support for subsequent status classification and trend prediction.

[0056] In this embodiment of the invention, the method for generating a dynamic handover strategy includes:

[0057] Obtain the output status data of the first conveyor line section and the input status data of the second conveyor line section, wherein the first conveyor line section and the second conveyor line section are adjacent conveyor line sections;

[0058] Based on the output status data and the input status data, calculate the probability of a handover conflict between the first conveyor line segment and the second conveyor line segment;

[0059] Based on the handover conflict probability, the handover priority of the first and second conveyor line segments is generated through a game theory model.

[0060] Based on the handover priority and the high-level business logic of the business host, a dynamic handover strategy is generated. The dynamic handover strategy includes the handover time window, handover speed, and handover path selection.

[0061] In this embodiment, the status data of adjacent conveyor line sections are first acquired in real time. For the first conveyor line section (upstream), its output status data is acquired, including: output product queue status (such as the number of products waiting to be handed over, product type, product size, etc.), output rate (such as the current number of products output per minute), output equipment status (such as the operating status of the drive device, the detection status of the sensor, etc.), product attributes (such as weight, shape, fragility, etc.), and the estimated time series of arrival at the handover point, etc. For the second conveyor line section (downstream), its input status data is acquired, including: input buffer status (such as current occupancy rate, remaining capacity, etc.), input rate (such as the current number of products that can be received per minute), input equipment status (such as the receiving device preparation status, the sensor ready status, etc.), processing capacity constraints (such as the maximum processable product size, weight limit, etc.), and the priority of the current processing task, etc. The status data is collected and updated in real time through the network communication module to ensure that decisions are based on the latest information.

[0062] Based on the acquired output and input state data, the probability of potential handover conflicts between the first and second conveyor line segments is calculated. Conflict types include: time conflicts (e.g., multiple products arriving at the handover point simultaneously), capacity conflicts (e.g., product parameters exceeding downstream processing capacity), and rate conflicts (e.g., upstream output rate exceeding downstream receiving rate). The conflict probability calculation employs a probabilistic model, such as a Bayesian network or a Markov model, considering various possible state combinations and their probabilities of occurrence. For example, the probability of a time conflict can be expressed as: P(time conflict) = P(|T1-T2|<Δt), where T1 and T2 are the arrival times of the two products at the handover point, and Δt is the safety time interval. By combining the model parameters trained using historical data with current state data, a real-time handover conflict probability matrix is ​​calculated, quantifying the likelihood of different types of conflicts occurring.

[0063] Based on the probability of handover conflicts, a game theory model is used to calculate the handover priority of the first and second conveyor line segments. The game theory model treats the handover process as a multi-partner strategic interaction, with each conveyor line segment choosing its optimal strategy based on its own state and objectives (such as maximizing throughput, minimizing waiting time, and optimizing energy consumption). The model employs a non-cooperative game framework, solving for the optimal strategy combination through Nash equilibrium. Priority calculation considers multiple factors, such as product urgency (based on delivery time), product value (based on product type and customer level), and system load (based on current processing volume). Priorities are represented numerically, with higher values ​​indicating higher priority. In the event of a conflict, higher-priority handover requests will receive priority processing, while lower-priority requests may be delayed or rescheduled.

[0064] Based on the calculated handover priority and the high-level business logic provided by the business host (such as order priority, delivery time requirements, capacity balancing strategies, etc.), a complete dynamic handover strategy is generated. The handover strategy comprises three key components: a handover time window, which clearly defines the optimal handover time period for each product, including start time, end time, and optimal handover moment, reducing conflicts through time staggering; a handover speed, which specifies the optimal operating speed for upstream and downstream equipment to achieve speed matching and smooth transition, avoiding product backlog or equipment idling due to speed mismatch; and a handover path selection, which, in a multi-path handover system, specifies the optimal handover path, considering factors such as path load, distance, and equipment status to achieve load balancing and optimal resource utilization. The dynamic handover strategy is distributed to relevant conveyor line sections via network communication modules to guide the execution of the actual handover process. The strategy execution effect is evaluated in real time through a feedback mechanism; if deviations occur, strategy adjustments are triggered, forming a closed-loop control to ensure the stability and efficiency of the handover process.

[0065] In this embodiment of the invention, the method for dynamically binding product tracking data with the physical location of the product includes:

[0066] Obtain the identification code of each physical product in the conveyor line section. The identification code is used to uniquely identify the physical product.

[0067] Based on the real-time position data of the physical products in the conveyor line section collected by the sensors, a virtual position mapping table of the physical products is generated.

[0068] Based on a virtual location mapping table, an interpolation algorithm is used to predict the continuous location trajectory of physical products in the conveyor line section.

[0069] The identification code is bound to a continuous location trajectory to generate product tracking data;

[0070] Product tracking data is stored in a distributed tracking database, which uses a sharded storage structure and manages the product tracking data in partitions according to the geographical location of the conveyor line sections.

[0071] In this embodiment, each physical product in the conveyor line section is first uniquely identified. Identification methods include: barcode labels, affixed to the product surface and read by a barcode scanner; RFID tags, embedded or affixed to the product and identified non-contactly by an RFID reader; QR code labels, printed or affixed to the product and read by an image recognition device; and visual feature recognition, utilizing the product's own appearance characteristics (such as shape, color, texture, etc.) through a machine vision system. The identification process is completed automatically when the product enters the conveyor system. Each product is assigned a unique identification code, such as a UUID (Universally Unique Identifier), serial number, or order number. The identification code is associated with the product's business attributes (such as category, specifications, order information, etc.) and stored in the system database, providing a basis for subsequent tracking.

[0072] Based on data collected by various sensors distributed along the conveyor line, the real-time position of physical products on the conveyor line is determined. Sensor types include: photoelectric sensors, which detect the time it takes for a product to pass a specific point; vision sensors, which determine product position through image analysis; RFID readers, which identify product positions within their reading range; and laser rangefinders, which accurately measure the distance between the product and a reference point. These sensors are distributed at certain intervals along the conveyor line, forming a sensor network. When a product passes through the sensor coverage area, the sensor triggers and records the product ID and timestamp, uploading this information to the control system. The system integrates the data from each sensor to construct a discrete set of product positions on the conveyor line, and combines this with the conveyor line's topology and coordinate system to generate a virtual position mapping table for the products. This mapping table is a dynamically updated data structure containing information such as product ID, timestamp, coordinate position, and movement status, reflecting the real-time distribution of products in physical space.

[0073] Based on discrete position data in a virtual position mapping table, interpolation algorithms are used to predict the continuous position trajectory of a product within a conveyor line segment. Interpolation algorithm types include: linear interpolation, suitable for simple scenarios with constant speed movement; spline interpolation, suitable for scenarios with smooth speed changes, generating smooth curves; Bézier curve interpolation, suitable for accurate descriptions of complex paths; and Kalman filtering, combining motion models and measurement data, suitable for trajectory prediction in noisy environments. The algorithm selection is dynamically determined based on the characteristics of the conveyor line and accuracy requirements. The interpolation process considers the physical constraints of the conveyor line (such as speed limits, acceleration limits, path constraints, etc.) and real-time operating parameters (such as current speed, equipment status, etc.) to ensure the physical feasibility of the predicted trajectory. The prediction result is a continuous time-position function, which can calculate the expected position of the product at any given time, compensating for the position information gaps caused by discrete sensor sampling and enabling continuous tracking of the product's position.

[0074] By binding product identification codes to predicted continuous location trajectories, complete product tracking data is generated. This tracking data is a multi-dimensional data structure containing: basic product information (such as ID, type, specifications, etc.), location trajectory information (such as current location, historical trajectory, predicted trajectory, etc.), time information (such as entry time, estimated completion time, and time to each node), status information (such as normal movement, paused, abnormal, etc.), and business information (such as order, priority, destination, etc.). This information is organized through relationships to form a complete digital mapping of the product, supporting comprehensive product tracking and monitoring.

[0075] The generated product tracking data is stored in a distributed tracking database to ensure high availability and efficient querying. The database employs a sharded storage structure, partitioning the tracking data according to the geographical location of the delivery line sections. Each physical region corresponds to a data shard, managed by the nearest server node, reducing data transmission latency. Shards maintain data consistency through a synchronization mechanism, ensuring the continuity of tracking data when products move across regions. The database supports multiple query modes, such as querying by product ID, by location region, and by time range, meeting the needs of different business scenarios. It also implements data lifecycle management, archiving or cleaning up historical data to maintain efficient system operation. This distributed architecture achieves efficient storage and fast access to product tracking data, providing reliable data support for upper-layer applications.

[0076] In this embodiment of the invention, the method for generating the prediction model includes:

[0077] Acquire historical operating data and historical status data of the conveyor line section. Historical operating data includes the operating parameters of the equipment objects, and historical status data includes the fault records and handover records of the equipment objects.

[0078] A time-series dataset is constructed based on historical operational data and historical status data;

[0079] Feature extraction is performed on the time series dataset to generate a prediction feature set, which includes trend features, periodic features, and anomaly features of the running parameters.

[0080] Based on the prediction feature set, a prediction model is trained through a long short-term memory network to generate prediction data of operating trends.

[0081] In this embodiment, comprehensive historical data for the conveyor line section is first obtained from the system's historical database. Historical operational data includes records of various parameters during the long-term operation of the equipment, such as historical speed curves of drive motors, current change records, temperature fluctuation data, historical sensor readings, trigger records, sensitivity changes, actuator action timing records, response time statistics, and torque output history. Historical status data includes equipment fault records, such as fault type, occurrence time, duration, severity, and handling method; handover records, such as handover time point, type of product handed over, handover success rate, and type of handover anomaly; and auxiliary information such as maintenance records and warning records. During data acquisition, a quality assessment is performed, checking the data's completeness, accuracy, and consistency, eliminating or repairing low-quality data to ensure the reliability of subsequent analysis.

[0082] Based on the acquired historical operational and status data, a standardized time-series dataset was constructed. The dataset is organized in a multi-dimensional time-series format, with each time point corresponding to multiple feature dimensions, including original operational parameters, status markers, and derived indicators. Data preprocessing includes: time alignment, unifying data from different sources and with different sampling rates onto the same time scale using interpolation or downsampling methods; missing value handling, filling data gaps using forward imputation, mean imputation, or model prediction to ensure data continuity; outlier handling, identifying and processing outliers in the data to avoid their negative impact on model training; and data standardization, normalizing parameters of different dimensions to make them suitable for model training. The processed dataset is arranged in chronological order and divided into training, validation, and test sets for model training and evaluation.

[0083] Feature extraction is performed on the constructed time-series dataset to uncover patterns and regularities, generating a predictive feature set. Feature extraction focuses on three key types of features: trend features, reflecting the long-term direction of parameter changes (e.g., linear trend coefficients, exponential smoothing values, moving averages, etc.), capturing the gradual change process of parameters; periodic features, reflecting the cyclical change patterns of parameters (e.g., Fourier transform coefficients, wavelet transform features, autocorrelation coefficients, etc.), identifying periodic patterns in the data; and anomaly features, reflecting abnormal fluctuations in parameters (e.g., mutation detection indicators, entropy changes, outlier measures), discovering potential precursors of anomalies. The feature extraction process employs a sliding window mechanism, calculating feature values ​​at different time scales to capture short-term, medium-term, and long-term data characteristics. Furthermore, interaction terms and derived features between features are introduced, such as correlation coefficients, ratios, and composite indicators between parameters, enhancing the expressive power of the features. After importance evaluation and redundancy analysis, the extracted features are selected to form the most predictive subset, constituting the final predictive feature set.

[0084] Based on the prediction feature set, a prediction model is constructed using a Long Short-Term Memory (LSTM) network from deep learning. The LSTM network structure includes: an input layer that receives feature vectors from the prediction feature set; an LSTM layer containing multiple LSTM units, each with an input gate, a forget gate, and an output gate, capable of learning long-term dependencies; an attention mechanism layer that enhances attention to key time points and key features to improve prediction accuracy; a fully connected layer that maps the output of the LSTM layer to the prediction target space; and an output layer that generates the final prediction result. The network is trained using the backpropagation algorithm, with the loss function chosen as either Mean Squared Error (MSE) or Mean Absolute Error (MAE), and the optimizer selected as either Adam or RMSprop. A dynamic learning rate adjustment strategy is employed. To prevent overfitting, regularization techniques such as L2 regularization and Dropout are introduced. During model training, model performance is evaluated using a validation set, and hyperparameters, such as the number of hidden layers, the number of units, and the learning rate, are dynamically adjusted to find the optimal configuration. After training, the model is evaluated using a test set to ensure good generalization ability.

[0085] The finally trained LSTM prediction model can generate operational trend prediction data for conveyor line sections based on current and historical operational data. The prediction data includes multiple dimensions: equipment performance trends, predicting future performance changes such as efficiency, energy consumption, and response time; fault risk prediction, assessing the probability and possible time windows of various equipment faults; handover capacity prediction, predicting the maximum handover capacity and potential bottlenecks in future periods; and resource demand prediction, estimating the energy and maintenance resources required for future operation. The prediction results are presented in time series format, including predicted values, confidence intervals, and possible scenario branches, providing multi-dimensional reference information for system control decisions. After deployment, the model employs an online learning mechanism to continuously absorb new operational data, continuously optimize model parameters, maintain the accuracy and timeliness of predictions, and adapt to the dynamic changes of the conveyor system.

[0086] In this embodiment of the invention, the dynamic power allocation method for the power management module of the modular smart driver card includes:

[0087] Obtain the real-time power requirements of each device object in the conveyor line section;

[0088] Based on real-time power demand, the peak power demand of the transmission line section within a future time window is predicted using a power prediction algorithm.

[0089] Based on the peak power demand, the power allocation ratio of the power management module is dynamically adjusted to generate a power allocation strategy.

[0090] Based on the power allocation strategy, the control power management module provides differentiated power outputs to the device objects.

[0091] In this embodiment, the power demand data of each device in the conveyor line section is first acquired in real time. For drive motors, real-time current, voltage, and power factor are measured to calculate active power, reactive power, and apparent power. For sensors, supply voltage and operating current are recorded to calculate power consumption levels. For actuators, instantaneous power and average power during operation are monitored. Power data acquisition employs a high-precision power parameter measurement circuit, supporting wide-range power monitoring (e.g., from 0.1W to several kW) and fast response (e.g., millisecond-level sampling rate). During data acquisition, the operating status and load conditions of the equipment are simultaneously recorded to establish a mapping relationship between power demand and operating status, providing a basis for subsequent prediction. After preliminary processing (e.g., filtering and calibration), all power data forms a real-time power demand data stream, which is then input into the power management system.

[0092] Based on acquired real-time power demand data, a power prediction algorithm is applied to predict power demand changes, particularly peak power, within future time windows (e.g., 5 minutes, 15 minutes, 30 minutes, etc.) for transmission line sections. The prediction algorithm integrates multiple techniques: time series forecasting methods, such as Autoregressive Moving Average (ARIMA) and exponential smoothing, predict based on the temporal patterns of historical power data; machine learning methods, such as Support Vector Regression (SVR) and Random Forest, utilize multi-dimensional features (including current power, equipment status, and product characteristics) to predict future power; and deep learning methods, such as Recurrent Neural Networks (RNN) and Long Short-Term Memory Networks (LSTM), capture complex nonlinear patterns of power changes. The prediction process considers business planning factors, such as the product queues to be processed and planned speed adjustments, enhancing the predictive foresight. The algorithm output includes expected power values ​​at different time points, power change curves, peak occurrence time and duration, and prediction confidence intervals, comprehensively describing the characteristics of future power demand.

[0093] Based on power demand forecasts, particularly the predicted peak power demand, the power allocation strategy of the power management module is dynamically adjusted. The strategy formulation process considers multiple factors: total power constraints, setting a power allocation upper limit based on the total capacity of the power system; equipment priority, determining power supply priorities based on equipment importance and business impact; energy efficiency optimization, maximizing the overall energy utilization efficiency of the system; and peak smoothing, balancing power peaks and valleys through pre-allocation and delayed power supply. The strategy is represented by a set of power allocation ratio coefficients, defining the percentage of power each equipment category or functional unit should receive under different load conditions. The allocation strategy is adaptive, dynamically adjusting according to real-time load changes and forecast results to ensure stable operation and optimal performance of the system under various operating conditions.

[0094] Based on the generated power allocation strategy, the power management module provides differentiated power output to each device in real time. Control methods include: voltage regulation, dynamically adjusting the output voltage through efficient voltage regulator circuits to meet the voltage requirements of different devices; current limiting, limiting the output current of each circuit through precise current control circuits to avoid overload; power modulation, adjusting power output through pulse width modulation (PWM) or phase control technology to achieve precise power allocation; and timing control, controlling the power-on time and sequence of each device through intelligent timing circuits to avoid power peaks caused by simultaneous startup. The power supply control features high precision (e.g., power control accuracy better than 1%) and fast response characteristics (e.g., millisecond-level adjustment speed), adapting to rapidly changing load conditions. Simultaneously, the control system continuously monitors the power supply status and device response, forming a closed-loop control to ensure the effective execution of the power allocation strategy. In extreme cases (e.g., total power demand exceeding system capacity), automatic protection mechanisms are activated, such as priority-based power cut-off and load degradation, to protect system safety. Through this intelligent dynamic power management, the system can achieve optimal energy allocation with limited power resources, improving energy utilization efficiency, reducing operating costs, ensuring the reliable operation of critical equipment, and enhancing overall system performance.

[0095] In this embodiment of the invention, the system further includes an adaptive configuration unit, which is used to adaptively configure parameters through a network interface. The method for adaptive parameter configuration includes:

[0096] Obtain the hardware parameters of the modular intelligent drive card and the operating parameters of the conveyor line section;

[0097] Based on hardware and operating parameters, the environmental perception algorithm identifies the operating environment characteristics of the conveyor line section, including temperature, humidity, load change rate, and equipment aging degree.

[0098] Based on the characteristics of the operating environment, an adaptive parameter configuration strategy is generated through a reinforcement learning model.

[0099] Based on an adaptive parameter configuration strategy, the operating parameters of the modular intelligent driver card and the main control unit are updated, and the update results are displayed through a network interface.

[0100] In this embodiment, the adaptive configuration unit, as the core configuration management component of the system, first acquires the hardware parameters of the modular intelligent driver card and the operating parameters of the transmission line segment through the system bus and network interface. Hardware parameters include: processor specifications (e.g., model, frequency, number of cores), memory capacity, communication interface type (e.g., RS485, CAN, Ethernet), power specifications (e.g., input voltage range, maximum power), expansion capabilities (e.g., number of I / O points, module slots), and hardware version number. Operating parameters include: processor load rate, memory utilization, communication bandwidth usage, power output, I / O response time, module temperature, and other real-time performance indicators. The parameter acquisition process employs a combination of polling and event triggering to ensure that the data is both comprehensive and timely. After preprocessing (e.g., noise reduction and formatting), the acquired parameters form a structured parameter dataset, providing a foundation for subsequent analysis.

[0101] Based on the acquired hardware and operational parameters, an environmental perception algorithm identifies the operating environment characteristics of the conveyor line section. This algorithm integrates multi-source data, including: direct environmental sensing data, such as physical environmental parameters like temperature, humidity, dust concentration, and vibration intensity collected by built-in or external sensors; indirect inference data, such as environmental factors inferred from changes in equipment operating characteristics, like power fluctuations and communication quality changes; and historical comparison data, comparing current operating parameters with historical baselines to identify abnormal change patterns. The algorithm employs multimodal data fusion technology to integrate and analyze data from different sources and in different formats to extract environmental features. Key environmental features include: temperature characteristics, including average temperature, temperature fluctuation range, and temperature gradient, affecting the reliability and lifespan of electronic components; humidity characteristics, including relative humidity and humidity change rate, affecting insulation performance and electrical safety; load change rate, characterizing the dynamic characteristics of the system load, such as peak-to-valley ratio, change frequency, and abrupt change amplitude, affecting system resource allocation strategies; and equipment aging degree, quantified through performance degradation curves, error growth rate, and response latency indicators, reflecting the health status and remaining lifespan of the equipment. These characteristics comprehensively reflect the operating environment of the conveyor line section, providing a basis for decision-making on adaptive parameter configuration.

[0102] Based on the identified operating environment characteristics, an adaptive parameter configuration strategy is generated using a reinforcement learning model. The reinforcement learning model models the parameter configuration problem as a Markov Decision Process (MDP), including: a state space, jointly defined by the operating environment characteristics and the current system state; an action space, including the value range and adjustment step size of each adjustable parameter; a reward function, designed based on system performance indicators (such as processing efficiency, energy consumption, failure rate, etc.), quantifying the effect of parameter adjustment; and a state transition model, describing the mechanism by which parameter adjustment affects the system state. The model employs advanced reinforcement learning algorithms, such as Deep Q-Networks (DQN), Policy Gradient, or Actor-Critic architecture, to learn the optimal parameter configuration strategy through interaction with the environment. During the learning process, the model gradually explores the parameter space, discovers the optimal parameter combinations under different environmental conditions, and establishes a mapping relationship between environmental characteristics and parameter configuration. Model training combines online and offline learning, utilizing historical data for basic training while continuously optimizing based on real-time feedback to adapt to dynamic environmental changes. The learning outcome is an adaptive parameter configuration strategy that can automatically recommend the optimal parameter settings based on the characteristics of the current operating environment, thereby achieving environmental adaptation of system performance.

[0103] Based on the generated adaptive parameter configuration strategy, the operating parameters of the modular intelligent driver card and the main control unit are updated. The parameter update process includes: parameter verification, which checks the legality and security of recommended parameters to ensure they are within safe limits; incremental updates, which employ a gradual adjustment strategy for key parameters to avoid system instability caused by sudden changes; a rollback mechanism, which sets up a monitoring window for parameter updates so that performance anomalies can be quickly rolled back to a previous stable configuration; and version management, which records historical versions of parameter updates and supports configuration comparison and rollback operations. After the parameter update, the system intuitively displays the update results through a web interface, including: a parameter change list, clearly showing the comparison of parameters before and after the adjustment; performance impact analysis, which quantifies the expected impact of parameter adjustments on various system performance indicators; environmental adaptability scoring, which assesses the adaptability of the current configuration to the current environment; and optimization suggestions, which provide directions and room for further optimization. The web interface adopts a responsive design and supports access from various terminal devices (such as PCs, tablets, and mobile devices). Interface elements include data tables, trend charts, parameter relationship network diagrams, etc., comprehensively and intuitively presenting configuration information. Through this adaptive parameter configuration mechanism, the system can proactively adjust its operating parameters according to environmental changes, maintain optimal operating conditions, improve the system's adaptability and reliability in complex and ever-changing environments, and reduce the workload and error risk of manual configuration.

[0104] In this embodiment of the invention, the method for generating the game theory model includes:

[0105] Construct a game objective function with the goal of minimizing the probability of handover conflict, and use the task priority of the high-level business logic of the business host as a constraint.

[0106] The handover behavior of the first and second conveyor line sections is modeled strategically to generate a handover strategy set.

[0107] Based on the game objective function and the set of handover strategies, the handover priority is solved using the Nash equilibrium algorithm.

[0108] In this embodiment, the objective function of the handover process is first constructed as the core evaluation criterion of the game model. The objective function design is based on two key factors: first, minimizing the probability of handover conflicts, using the probability of various conflicts (such as time conflicts, capability conflicts, rate conflicts, etc.) as the main optimization objective to minimize the overall system conflict rate; second, using the task priorities in the high-level business logic provided by the business host as constraints to ensure that high-priority tasks are given priority. The mathematical expression of the objective function is a multi-objective optimization function, which can be represented as:

[0109] F(x)=α·P_conflict(x)+β·∑ i w_i·V_i(x);

[0110] Where P_conflict(x) represents the overall conflict probability under policy x; V_i(x) represents the impact of policy x on the task with priority i (such as delay, cost, etc.); w_i represents the weight coefficient of priority i; α and β are balancing coefficients used to adjust the balance between conflict minimization and priority protection. The objective function may also include other constraints, such as energy consumption limits and equipment load balancing, forming a complete multi-objective optimization problem.

[0111] Strategy modeling is performed on the handover behavior between the first and second conveyor line sections, abstracting the possible actions and decisions of each participant into a strategy space. The strategy space for the first conveyor line section (upstream) includes: output rate adjustment strategies, such as maintaining the current rate, increasing the rate, and decreasing the rate; product sorting strategies, such as sorting by product priority, sorting by product attribute, and sorting by destination; buffer management strategies, such as active buffering, direct output, and conditional buffering; and handover request strategies, such as immediate request, delayed request, and batch request. The strategy space for the second conveyor line section (downstream) includes: receiving rate adjustment strategies, such as matching the upstream rate, fixed rate, and dynamically adjusting the rate; admission control strategies, such as full reception, conditional reception, and rejection; buffer management strategies, such as reserved buffer, dynamic allocation, and priority buffering; and handover response strategies, such as immediate response, delayed response, and conditional response. These strategies can be combined to form composite strategies, constituting a complete strategy set. The strategy modeling process considers equipment physical constraints, control precision limitations, and response time characteristics to ensure the practical feasibility of the model.

[0112] Based on the constructed game objective function and handover strategy set, the Nash equilibrium algorithm is applied to solve for the handover priority. The solution process first establishes a payoff matrix for the strategy combinations, where each element represents the payoff (or cost) for each participant under a specific strategy combination, such as the degree of conflict probability reduction or task completion efficiency improvement. The payoff matrix may be high-dimensional, containing multiple strategy combinations and multiple participants. The Nash equilibrium solution employs iterative methods, such as Best Response Dynamics, Pseudo-gradient Descent, or Evolutionary Game Algorithms. Constraints, such as physical feasibility constraints and resource limitations, are considered during the solution process to ensure the result is executable in a real system. The algorithm outputs multiple possible equilibrium points, each corresponding to a set of handover strategy combinations and a corresponding priority allocation scheme. By evaluating the performance indicators of each equilibrium point (such as conflict probability, system throughput, task latency, etc.), the optimal equilibrium solution is selected as the final handover priority scheme. The scheme represents the priority order of each handover request in numerical form, such as a priority score based on 0-1 normalization, which facilitates the system's priority judgment and resource allocation. The handover priority solution process is dynamic and is updated in real time as the system state and business requirements change, ensuring that the priority always reflects the current optimal decision.

[0113] In this embodiment of the invention, the method for predicting the continuous position trajectory of a physical product in a conveyor line segment using an interpolation algorithm includes:

[0114] Obtain discrete location data of physical products from the virtual location mapping table. The discrete location data includes timestamps and corresponding virtual location coordinates.

[0115] Based on the virtual location mapping table, the location change characteristics of physical products in the conveyor line section are extracted. The location change characteristics include the discrete time interval of the location coordinates and the location offset.

[0116] Based on the characteristics of position change, a dynamic position distribution model of physical products in the conveyor line section is constructed. The dynamic position distribution model is used to characterize the position distribution probability of physical products in different time periods.

[0117] Based on the dynamic position distribution model, a probability-weighted interpolation algorithm is used to interpolate discrete position data to generate a preliminary continuous position trajectory of physical products in the conveyor line section.

[0118] Acquire the operating status data of the equipment objects in the conveyor line section. The operating status data includes the speed fluctuation of the drive motor and the detection error of the sensor.

[0119] Based on the operating status data, the initial continuous position trajectory is corrected to generate a corrected continuous position trajectory. The correction includes adjusting the smoothness of the position trajectory based on the speed fluctuation and adjusting the offset of the position trajectory based on the detection error.

[0120] The corrected continuous location trajectory is dynamically updated with the virtual location mapping table to ensure that the virtual location mapping table is consistent with the actual location of the physical product.

[0121] In this embodiment, discrete location data of the physical products are first extracted from a virtual location mapping table. The location mapping table is a dynamic data structure maintained by the system, recording the movement trajectory points of each product on the conveyor line. The discrete location data contains two key elements: a timestamp, which precisely records the moment the product is detected, typically with millisecond-level accuracy; and virtual location coordinates, which record the product's position in a virtual coordinate system. This coordinate system can be one-dimensional (e.g., distance values ​​on a straight conveyor line), two-dimensional (e.g., (x,y) coordinates in a planar conveyor system), or three-dimensional (e.g., (x,y,z) coordinates in a three-dimensional conveyor system). The data extraction process includes a data quality check, filtering outliers (e.g., points that significantly deviate from a reasonable trajectory) and redundant data (e.g., repeated sampling points with excessively short time intervals) to ensure the data quality for subsequent analysis. The extracted discrete location data forms a time-location sequence, providing a basis for trajectory prediction.

[0122] Based on historical data in the virtual location mapping table, the positional change characteristics of physical products within the conveyor line segment are extracted. These characteristics comprise two core dimensions: discrete time intervals of position coordinates, reflecting the frequency and regularity of position sampling; and position offsets, representing the positional changes between adjacent sampling points. The vector characteristics (magnitude, direction) and statistical characteristics (average speed, acceleration, turning frequency) of these offsets are calculated. The feature extraction process employs time series analysis methods, such as sliding window statistics, rate of change calculation, and pattern recognition, to uncover motion patterns from discrete data. Furthermore, context-dependent features are extracted, such as behavioral characteristics in specific areas (e.g., turns, intersections, junctions) and motion characteristics under specific conditions (e.g., high load, high speed operation). These positional change characteristics comprehensively describe the motion characteristics of the product on the conveyor line, providing a basis for subsequent modeling.

[0123] Based on the extracted position change features, a dynamic position distribution model of physical products in the conveyor line section is constructed. This model is a probabilistic description of the product's position change over time, capable of characterizing the possible position distribution of the product at any given moment. The model is constructed using probabilistic statistical methods, such as Gaussian processes, Hidden Markov models, or Bayesian networks. Model parameters are obtained through training on historical data, capturing both the regularity and randomness of product movement. Model characteristics include: time-varying nature (model parameters dynamically adjust over time to adapt to changes in system state); spatial correlation (considering the correlation structure between positions, such as path constraints and velocity continuity); and multimodal nature (capable of handling various possible motion modes, such as uniform motion, accelerated motion, and pauses). The model output is the probability density function of the product's position distribution over different time periods, providing both the most probable position and quantifying the uncertainty of position prediction.

[0124] Based on the constructed dynamic position distribution model, a probabilistic weighted interpolation algorithm is applied to interpolate discrete position data to generate preliminary continuous position trajectories. Unlike traditional deterministic interpolation methods (such as linear interpolation and spline interpolation), the probabilistic weighted interpolation algorithm considers the probability distribution of position prediction, thus more accurately reflecting the uncertainty of product movement. The algorithm flow includes: calculating the probability distribution of possible positions for each time point to be interpolated based on the dynamic position distribution model; sampling multiple possible position points from the probability distribution; performing a weighted average of the sampled points, with weights based on the probability value and physical feasibility of the position; and generating a smooth trajectory that satisfies physical constraints (such as speed limits, acceleration limits, path constraints, etc.). The interpolation process considers the topology of the conveyor line, such as straight segments, curved segments, and intersections, to ensure that the generated trajectory conforms to physical path constraints. The algorithm output is a time-continuous position trajectory function, which can calculate the expected position at any time, bridging the time gaps in the original discrete data.

[0125] Real-time operating status data of equipment within the conveyor line section is acquired to provide a basis for trajectory correction. The operating status data includes two key types of information: the speed fluctuation of the drive motor, reflecting changes in the conveyor line speed, including real-time speed values, speed fluctuation range, acceleration and deceleration characteristics, etc.; and sensor detection errors, quantifying the uncertainty of position detection, including systematic errors (such as sensor position offset, signal delay, etc.) and random errors (such as noise interference, environmental influences, etc.). The status data is acquired in real-time through the monitoring interface of the modular intelligent drive card. After data preprocessing (such as filtering and calibration), a structured status dataset is formed for trajectory correction.

[0126] Based on the acquired equipment operating status data, the initial continuous position trajectory is corrected to generate a more accurate continuous position trajectory. The correction process includes two key steps: first, adjusting the trajectory smoothness based on drive motor speed fluctuations, mapping actual motor speed changes (such as fluctuations, jitters, transients, etc.) onto the position trajectory to more realistically reflect the product's actual motion state (e.g., the trajectory becomes smoother when the motor decelerates and steeper when the motor accelerates); second, adjusting the trajectory offset based on sensor detection errors, compensating for position offsets caused by sensor error characteristics (such as system bias, random noise, etc.). The correction employs filtering algorithms such as Kalman filters or particle filters to fuse the initial trajectory with the status data, preserving the trajectory's continuity and smoothness while enhancing its accuracy and reliability. The correction process is dynamic; as new status data is continuously generated, the trajectory is continuously updated and optimized, always maintaining consistency with the actual product position.

[0127] The corrected continuous position trajectory and the virtual position mapping table are dynamically updated to ensure that the position data in the system always reflects the actual position of the product. The update process employs an incremental update strategy, updating only the changed parts to reduce system overhead. Updates include: fine-tuning of historical trajectory points, adjusting recorded historical points based on newly acquired information to improve the accuracy of historical data; precise positioning of the current location, calculating the product's current position in real time to support real-time monitoring and interaction; and predictive updates of future trajectories, adjusting future trajectory predictions based on the latest data to improve prediction accuracy. Data consistency checks are implemented during the update process to ensure that the updated data meets physical and logical constraints, such as velocity continuity and position boundary limitations. Through this dynamic update mechanism, the virtual position mapping table can be continuously optimized, constantly improving the accuracy and timeliness of product position representation, providing reliable position data support for product tracking and system control.

[0128] Through the above description of specific embodiments, the integrated control system for logistics conveyor lines based on modular intelligent drive cards of the present invention realizes real-time acquisition and analysis of equipment object operating parameters, dynamic classification and prediction of operating status, seamless handover and coordination between adjacent conveyor line sections, and accurate tracking and mapping of product positions, which significantly improves the intelligence level, operating efficiency and reliability of the logistics conveyor system.

[0129] The above description is merely a preferred embodiment of the present invention and is 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.

[0130] All formulas in this manual are dimensionless and calculated numerically. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0131] Although embodiments of the invention have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the claims and their equivalents.

Claims

1. An integrated control system for a logistics conveyor line based on modular intelligent drive cards, characterized in that, include: At least one modular intelligent drive card, at least one conveyor line section, at least one main control unit, and at least one service host; The modular smart drive card includes a power management module and a network communication module. The power management module is used to provide power, perform dynamic power allocation and collect operating status data. The network communication module is used to realize multi-protocol data communication and real-time handover control. The conveyor section includes at least one device object, which includes a drive motor, a sensor, and an actuator. The main control unit is used to communicate with the modular intelligent drive card and to control the conveyor line section in real time through a standardized communication interface; The standardized communication interface includes a vertical interface and a horizontal interface, wherein: The vertical interface includes a control interface and a reporting interface. The control interface is used to send optimization control commands from the main control unit to the modular intelligent drive card, and the reporting interface is used to upload the operating status data and operating feature vector from the modular intelligent drive card to the main control unit. The horizontal interface includes a dynamic handover interface, which is used to define the product handover protocol between adjacent conveyor line sections. The product handover protocol includes dynamic request signals, dynamic response signals, handover priority signals, and extended auxiliary signals. The data format of the dynamic handover interface adopts a variable-length data frame structure, which is used to carry product handover information, product tracking data, and priority information of the dynamic handover strategy. The business host is used to execute high-level business logic, including logistics task scheduling, data storage, and product tracking; The system achieves integrated control of the logistics conveyor line through the following steps: The operating parameters and status data of each device in the conveyor line section are obtained, and the corresponding operating feature vector is generated through the modular intelligent drive card. Based on the operating feature vector, the operating status of the conveyor line section is classified by a dynamic clustering algorithm to generate a set of classified operating statuses. Based on the set of operating states and the high-level business logic of the business host, predictive data of the operating trend of the conveyor line section is generated by a predictive model. Based on the predicted operating trend data, optimization control commands are generated and issued through the standardized communication interface to adjust the operating parameters of the device object. Based on the product handover requirements between the conveyor line sections, a dynamic handover strategy is generated through the network communication module, and the product handover between adjacent conveyor line sections is coordinated based on the dynamic handover strategy. The method for generating the dynamic handover strategy includes: Obtain the output status data of the first conveyor line section and the input status data of the second conveyor line section, wherein the first conveyor line section and the second conveyor line section are adjacent conveyor line sections; Based on the output status data and the input status data, calculate the probability of a handover conflict between the first conveyor line segment and the second conveyor line segment; Based on the handover conflict probability, the handover priority of the first conveyor line segment and the second conveyor line segment is generated through a game theory model; Based on the handover priority and the high-level business logic of the service host, the dynamic handover strategy is generated, which includes a handover time window, a handover speed, and a handover path selection. The method for generating the game theory model includes: Construct a game objective function, which takes minimizing the handover conflict probability as the optimization objective and takes the task priority of the high-level business logic of the business host as the constraint condition. The handover behavior between the first conveyor line segment and the second conveyor line segment is modeled using a strategy to generate a handover strategy set. Based on the game objective function and the set of handover strategies, the handover priority is solved using the Nash equilibrium algorithm. Based on the high-level business logic of the business host, product tracking data is generated, and the product tracking data is dynamically bound to the physical product location in the conveyor line section through a virtual mapping mechanism.

2. The integrated control system for logistics conveyor lines based on modular intelligent drive cards according to claim 1, characterized in that, The method for generating the running feature vector includes: The operating parameters of each device in the conveyor line section are obtained, including motor speed, sensor readings, and actuator status. The operating parameters are normalized to generate a set of normalized operating parameters; Based on the normalized set of operating parameters, the operating fluctuation characteristics of the device object within a preset time window are extracted. The operating fluctuation characteristics include the variance, frequency distribution, and peak change rate of the operating parameters. The operational fluctuation characteristics are fused with the status data to generate the operational feature vector of the device object.

3. The integrated control system for logistics conveyor lines based on modular intelligent drive cards according to claim 1, characterized in that, The method for dynamically binding product tracking data to the physical location of the product includes: Obtain the identification code of each physical product in the conveyor line section; the identification code is used to uniquely identify the physical product. Based on the real-time position data of the physical product in the conveyor line section collected by the sensor, a virtual position mapping table of the physical product is generated; Based on the virtual location mapping table, the continuous location trajectory of the physical product in the conveyor line section is predicted by an interpolation algorithm; The identification code is bound to the continuous location trajectory to generate the product tracking data; The product tracking data is stored in a distributed tracking database, which adopts a sharded storage structure and manages the product tracking data in partitions according to the geographical location of the conveyor line section.

4. The integrated control system for logistics conveyor lines based on modular intelligent drive cards according to claim 1, characterized in that, The method for generating the prediction model includes: Obtain historical operating data and historical status data of the conveyor line section. The historical operating data includes the operating parameters of the equipment object, and the historical status data includes the fault records and handover records of the equipment object. Based on the historical operational data and the historical state data, a time series dataset is constructed; Feature extraction is performed on the time series dataset to generate a prediction feature set, which includes trend features, periodic features, and anomaly features of the running parameters; Based on the predicted feature set, the prediction model is trained using a long short-term memory network to generate the predicted running trend data.

5. The integrated control system for logistics conveyor lines based on modular intelligent drive cards according to claim 1, characterized in that, The dynamic power allocation method of the power management module of the modular smart driver card includes: Obtain the real-time power requirement of each device in the conveyor line section; Based on the real-time power demand, the peak power demand of the transmission line section in a future time window is predicted using a power prediction algorithm. Based on the peak power demand, the power allocation ratio of the power management module is dynamically adjusted to generate a power allocation strategy; According to the power allocation strategy, the power management module is controlled to provide differentiated power output to the device.

6. The integrated control system for logistics conveyor lines based on modular intelligent drive cards according to claim 1, characterized in that, The system further includes an adaptive configuration unit, which is used to adaptively configure parameters through a network interface. The method for adaptively configuring the parameters includes: Obtain the hardware parameters of the modular intelligent drive card and the operating parameters of the conveyor line section; Based on the hardware parameters and the operating parameters, the operating environment characteristics of the conveyor line section are identified by an environmental perception algorithm. The operating environment characteristics include temperature, humidity, load change rate, and equipment aging degree. Based on the characteristics of the operating environment, an adaptive parameter configuration strategy is generated through a reinforcement learning model. Based on the adaptive parameter configuration strategy, the operating parameters of the modular intelligent driver card and the main control unit are updated, and the update results are displayed through the network interface.

7. The integrated control system for logistics conveyor lines based on modular intelligent drive cards according to claim 3, characterized in that, The method for predicting the continuous position trajectory of the physical product in the conveyor line segment using an interpolation algorithm includes: Obtain discrete location data of the physical products in the virtual location mapping table, wherein the discrete location data includes timestamps and corresponding virtual location coordinates; Based on the virtual location mapping table, the position change characteristics of the physical product in the conveyor line section are extracted, and the position change characteristics include the discrete time interval of the position coordinates and the position offset. Based on the position change characteristics, a dynamic position distribution model of the physical product in the conveyor line section is constructed; Based on the dynamic position distribution model, the discrete position data is interpolated using a probability weighted interpolation algorithm to generate a preliminary continuous position trajectory of the physical product in the conveyor line section. Obtain the operating status data of the equipment object in the conveyor line section, the operating status data including the speed fluctuation of the drive motor and the detection error of the sensor; Based on the operating status data, the preliminary continuous position trajectory is corrected to generate a corrected continuous position trajectory. The correction includes adjusting the smoothness of the position trajectory based on the rotational speed fluctuation and adjusting the offset of the position trajectory based on the detection error. The corrected continuous position trajectory is dynamically updated with the virtual position mapping table to ensure that the virtual position mapping table is consistent with the actual position of the physical product.

Images