Electric control system for logistics equipment

Abstract

The invention provides an electric control system for logistics equipment, and relates to the technical field of logistics system control, and the system comprises a data collection module which is used for collecting equipment state data and environment data in real time; the main control module is used for generating a scheduling strategy; the intelligent strategy optimization module is used for constructing a virtual model according to the equipment state data and the scheduling strategy; the plurality of logistics equipment ends are used for receiving and executing the control instruction; the fault feedback module is used for judging whether a fault occurs among the plurality of logistics equipment or not; and the standby line management module is used for judging whether standby equipment needs to be started or not. According to the invention, through a multi-layer high-precision dynamic interaction model constructed by the intelligent strategy optimization module, synchronous evolution of a virtual model and physical equipment is realized; and the fault feedback module accurately identifies a fault propagation path based on sliding window statistics and association rule mining, and realizes fault early warning through an LSTM prediction model in combination with an attention mechanism.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of logistics system control, and particularly relates to an electric control system for a logistics device. BACKGROUND

[0002] In the modern logistics industry, with the rapid development of e-commerce and the increasing complexity of global supply chain systems, the automation, intelligence and efficiency of logistics equipment have become key factors in promoting the sustainable development of the logistics industry. As the core control brain of logistics equipment, the electric control system is playing a crucial role. From the development history of logistics equipment, traditional logistics operations largely rely on manual handling and simple mechanical assistance. However, with the rapid growth of logistics business volume and the increasing demand for service quality, this traditional model gradually reveals its limitations such as low efficiency, high labor cost and prone to errors. In order to cope with these challenges, logistics equipment gradually evolves towards automation. Automated logistics equipment covers a wide range of types, such as automated guided vehicles (AGV), stackers, conveyors, etc. These devices can efficiently complete tasks such as cargo handling, storage and sorting in logistics warehouses, distribution centers and other places. The electric control system is the key to endowing these devices with "intelligence" and "action ability". The basic function of the electric control system is to accurately control the various moving parts of the logistics equipment, such as the speed, steering of the motor, and the action sequence of the actuator. Taking motor control as an example, through the electric control system, functions such as starting, stopping, accelerating, decelerating and speed regulation of the motor can be realized. This is the basis for ensuring that the logistics equipment can accurately run according to the predetermined trajectory and speed. For example, during the operation of AGV, the electric control system needs to accurately control the torque and speed of the motor according to the pre-planned path planning information, so that the AGV can smoothly travel in the aisle of the warehouse and accurately stop at the cargo loading and unloading point. The electric control system of logistics equipment integrates motor control, sensor technology, communication technology and flexible control capability, etc. It is the core supporting technology for the automation, intelligence and efficiency of modern logistics equipment, and has effectively promoted the development of the logistics industry towards higher efficiency and precision.

[0003] The existing electric control system of logistics equipment relies on static rules or simple heuristic algorithms to generate scheduling strategies, lacking comprehensive analysis and optimization capabilities based on multi-dimensional data. The generation of task priority queues does not combine feature extraction and consistency checking mechanisms, which can easily lead to scheduling conflicts or resource waste. At the same time, traditional strategies cannot dynamically adjust according to real-time data, making it difficult to adapt to the high dynamics and complexity of the logistics environment, resulting in low equipment utilization and limited scheduling efficiency. Moreover, existing systems rely on post-hoc manual investigation or simple threshold alarms, lacking intelligent diagnostic capabilities based on historical fault knowledge base and machine learning models. Fault propagation path identification relies on human experience, making it difficult to quickly locate the root cause and block the propagation chain.

[0004] In order to solve the defects in the prior art, the technical scheme provides an electric control system for a logistics device. SUMMARY

[0005] The present application provides an electric control system for a logistics device to solve the defects in the prior art.

[0006] In one aspect, the present application provides an electric control system for a logistics device, comprising: a data acquisition module for acquiring device state data and environmental data in real time; a main control module for generating a scheduling strategy based on the device state data and the environmental data; an intelligent strategy optimization module for constructing a virtual model based on the device state data and the scheduling strategy, and outputting a control instruction; a plurality of logistics device terminals for receiving and executing the control instruction, and outputting an execution log; a fault feedback module for determining whether a fault occurs between the plurality of logistics devices based on the execution log, and outputting feedback information; a backup line management module for determining whether a backup device needs to be enabled based on the device state data and the feedback information, and outputting a backup line scheduling strategy if needed.

[0007] According to the electric control system for a logistics device provided by the present application, the main control module includes a task analysis unit and a strategy generation unit. The task analysis unit is used to formulate a task priority queue and a constraint condition set based on the device state data and the environmental data. The strategy generation unit is used to generate a scheduling strategy based on the task priority queue and the constraint condition set based on an operations research algorithm.

[0008] According to the electric control system for a logistics device provided by the present application, the steps of the task analysis unit formulating the task priority queue and the constraint condition set include: performing data cleaning, feature extraction and standardization processing on the device state data and the environmental data to construct a multi-dimensional data set containing device operating parameters, task space-time attributes and environmental dynamic indicators; calculating a task priority value using a multi-level evaluation model based on the multi-dimensional data set to generate a task priority queue; establishing a hierarchical constraint system based on the device operating parameters and the environmental dynamic indicators to output an initial constraint condition set; detecting conflicts in the task priority queue and the initial constraint condition set using a finite state machine model, dynamically adjusting non-critical constraint parameters through a constraint relaxation algorithm, and outputting a constraint condition set.

[0009] According to the application, a kind of electric control system for logistics equipment is provided, and intelligent strategy optimization module includes model construction unit, simulation test unit and instruction conversion unit;Model construction unit is used to construct logistics equipment high-precision virtual model according to equipment state data and scheduling strategy;Simulation test unit is used to simulate the execution scheduling strategy by logistics equipment high-precision virtual model and environmental data, generates simulation report;Instruction conversion unit is used to adjust scheduling weight according to simulation report, and generates control instruction according to scheduling weight.

[0010] According to the application, a kind of electric control system for logistics equipment is provided, and the steps that model construction unit constructs logistics equipment high-precision virtual model include: Preprocess equipment state data and environmental data, output preprocessed model parameters; Based on equipment state data and task demand, determine multi-layer modeling architecture, obtain virtual model; Preprocessed model parameters are input into virtual model, and model variables of virtual model are calibrated by historical operation record, to verify the behavior consistency of virtual model and logistics equipment; Integrate the subsystem model of mechanical motion, energy consumption and sensor feedback, and construct high-precision dynamic interaction model; Construct real-time data interaction channel of virtual model and high-precision dynamic interaction model and sensor and control system, and output logistics equipment high-precision virtual model.

[0011] According to the application, a kind of electric control system for logistics equipment is provided, and logistics equipment end includes instruction analysis unit and instruction execution unit, and instruction analysis unit is used to disassemble control instruction into bottom layer control parameter, and instruction execution unit is used to drive each device terminal according to bottom layer control parameter, and output execution log.

[0012] According to the application, a kind of electric control system for logistics equipment is provided, and the steps that instruction analysis unit disassembles control instruction into bottom layer control parameter include: Receive control instruction from upper layer system, parse instruction structure by ASCII / HEX protocol, extract instruction header, operation code, device identifier and target parameter field; According to operation code, match pre-defined instruction mapping table, convert high-level semantic instruction into device layer action combination, and output target parameter; Split target parameter into multiple logical parameters in multiple dimensions, including numerical parameter, coordinate parameter and time parameter;Numerical parameter is converted into PWM duty ratio by PID formula, coordinate parameter is converted into step motor step number by kinematics model, and time parameter is converted into timer interrupt period by RTC clock chip parameter; Cluster and integrate multiple logical parameters, and output bottom layer control parameter.

[0013] The electric control system for the logistics equipment provided by the application comprises a fault feedback module, a log analysis unit and a fault diagnosis unit.

[0014] The step of identifying the fault propagation path by the log analysis unit through the association rule mining comprises: constructing a device adjacency matrix to define the workflow dependency between devices; extracting key features of device operation anomalies based on the workflow dependency and the device collaborative anomaly; defining the item set of the association rule mining according to the key features and setting the rule parameters of the association rule mining; identifying the fault propagation path using the time sequence association rule in combination with the rule parameters.

[0015] The step of the fault diagnosis unit establishing the LSTM prediction model based on the historical fault knowledge base to give an early warning of the fault comprises: extracting the fault features in the historical fault indication library and performing feature statistics on the fault features according to a time window to output the time window features; dividing the time window features into a training set, a validation set and a test set in chronological order; constructing an LSTM model and training the LSTM model using the training set; verifying the LSTM model loss using the validation set, and terminating the training when the model loss does not decrease continuously for multiple rounds; checking the early warning timeliness of the LSTM model for the sudden fault on the test set.

[0016] The application provides an electric control system for a logistics device, a multi-layer high-precision dynamic interaction model constructed by adopting an intelligent strategy optimization module, a history data calibration and real-time data interaction channel, synchronous evolution of a virtual model and a physical device, a simulation test unit capable of predicting strategy execution effect and dynamically adjusting weights, generated control instructions subjected to multi-dimensional processing such as PID conversion and kinematics modeling, and ensured accuracy and stability of device action. The fault feedback module is based on sliding window statistics and association rule mining, accurately identifies the fault propagation path, and realizes fault warning through an LSTM prediction model combined with an attention mechanism. The deployment of a lightweight edge computing node and a federated learning framework supports low-latency response and multi-device collaborative evolution in time-sensitive scenarios, significantly reduces the average repair time and improves the system robustness. The standby line management module comprehensively evaluates the device state through a Markov decision process, combines the cold / hot standby mode and the digital twin pre-play mechanism, realizes seamless switching or delayed starting of the standby device. The design ensures the continuity of the system while optimizing the maintenance cost and fault risk, and the effectiveness of the emergency plan is further strengthened through simulation verification. BRIEF DESCRIPTION OF DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative labor.

[0018] Fig. 1 is a structural schematic diagram of an electric control system for a logistics device provided by an embodiment of the present application; Fig. 2 is a flowchart of constructing a high-precision virtual model of a logistics device in an electric control system for a logistics device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical scheme and advantages of the present application clearer, the technical scheme in the present application will be described clearly and completely in combination with the drawings in the present application. Obviously, the described embodiments are some embodiments of the present application, not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative labor belong to the protection scope of the present application.

[0020] Embodiment one: The following will describe an electric control system for a logistics device in combination with Figs. 1-2

[0021] As Fig. 1 ​As shown, the electric control system for the logistics equipment provided by the embodiment of the application comprises a data acquisition module, a main control module, an intelligent strategy optimization module, a plurality of logistics equipment ends, a fault feedback module and a standby line management module.

[0022] The data acquisition module is used for collecting equipment state data and environmental data in real time. The module adopts a distributed sensing network architecture, integrates multi-modal sensors such as temperature, vibration and position, and supports millisecond-level data sampling frequency. Data preprocessing is realized through edge computing nodes, including outlier rejection, time series alignment and spatial interpolation algorithms, effectively solving the data fusion problem caused by equipment heterogeneity. In view of the dynamic characteristics of the logistics environment, the digital twin technology is introduced to construct the equipment digital mirror, realizing real-time mapping of the physical space and the information space. The data acquisition module also has a data encryption function, which uses the AES encryption algorithm to encrypt the collected data, ensuring the security of the data in the transmission process. The AES algorithm is a symmetric encryption algorithm with high security and efficiency. The encryption process is as follows: Initialize the key to generate a random 128-bit key.

[0023] Group encryption, divide the data into 128-bit blocks, and encrypt each block separately.

[0024] Encryption formula: C = E(K, P), where C is the ciphertext, E is the encryption function, K is the key, and P is the plaintext.

[0025] Transmit the encrypted data to the main control module through the network.

[0026] The main control module is used to generate a scheduling strategy according to the equipment state data and the environmental data.

[0027] The main control module comprises a task analysis unit and a strategy generation unit, the task analysis unit is used to formulate a task priority queue and a constraint condition set according to the equipment state data and the environmental data. The steps of formulating the task priority queue and the constraint condition set comprise: Data cleaning, feature extraction and standardization processing are performed on the equipment state data and the environmental data to construct a multi-dimensional data set containing equipment operating parameters, task space-time attributes and environmental dynamic indicators. Wavelet transform is used to remove noise, and mutual information method is used for feature selection. The formula of wavelet transform is:

[0028] In the formula, W f (a, b) is the wavelet coefficient, f(t) is the initial data, i.e. the equipment state data and the environmental data. is the complex conjugate of 2(R) If the Fourier transform ψ(w) satisfies the admissibility condition:

[0029] ψ(t) is called a basic wavelet or mother wavelet. The mother wavelet ψ(t) is stretched and translated to obtain:

[0030] where a is a scale parameter and a>0, controlling the stretching of the wavelet. b is a translation parameter, controlling the translation of the wavelet.

[0031] Feature extraction requires the use of mutual information method. The formula is:

[0032] where X and Y are random data in device state data and environmental data, p(x,y) is the joint probability distribution, and p(x) and p(y) are the marginal probability distributions.

[0033] According to the multi-dimensional data set, a multi-level evaluation model is used to calculate the task priority value, and a task priority queue is generated, including: A hierarchical model is constructed, and the task is decomposed into target layer, criterion layer and scheme layer.

[0034] A judgment matrix is constructed according to the historical database score.

[0035] The eigenvalue method is used to calculate the weight, and the formula is:

[0036] where V is the characteristic weight, is the eigenvector, n is the order of the judgment matrix, that is, the number of rows and columns of the judgment matrix. i is the feature index.

[0037] The consistency ratio CR is calculated, and the formula is:

[0038]

[0039] where CI is the consistency index, RI is the random consistency index, is the maximum eigenvalue, which is obtained by calculation, where A is the judgment matrix and I is the unit matrix. For an n×n judgment matrix, the maximum eigenvalue is the largest one among all eigenvalues. If CR<0.1, it is considered that the consistency of the judgment matrix is within an acceptable range, and the subsequent weight calculation can continue. If CR≥0.1, it is considered that the consistency of the judgment matrix is poor, and the construction of the judgment matrix needs to be re-performed to improve the consistency.

[0040] Based on equipment operating parameters and environmental dynamic indicators, a hierarchical constraint system is established, and an initial constraint condition set is output.

[0041] A finite state machine model is used to detect conflicts in the task priority queue and initial constraint set. A constraint relaxation algorithm is then used to dynamically adjust non-critical constraint parameters, outputting the constraint set. The constraint relaxation algorithm can employ a linear programming approach with the following objective function:

[0042]

[0043] In the formula, minZ is the objective function, and c j For the target coefficient, sj Let α be the decision variables, i.e., the initial set of constraints. Let α be the constraint coefficients, and b be the initial set of constraints. j These are constraint values.

[0044] The strategy generation unit generates scheduling strategies based on operations research algorithms, taking into account task priority queues and constraint sets. The strategy generation unit also has a strategy optimization function, using reinforcement learning algorithms to optimize the generated scheduling strategies, thereby improving scheduling efficiency and equipment utilization.

[0045] The intelligent strategy optimization module is used to build a virtual model and output control commands based on device status data and scheduling strategies.

[0046] The intelligent strategy optimization module includes a model building unit, a simulation testing unit, and an instruction conversion unit. The model building unit constructs a high-precision virtual model of the logistics equipment based on equipment status data and scheduling strategies. The simulation testing unit simulates the execution of the scheduling strategy using the high-precision virtual model of the logistics equipment and environmental data, generating a simulation report. The instruction conversion unit adjusts the scheduling weights based on the simulation report and generates control instructions according to the scheduling weights.

[0047] like Fig. 2 As shown, the steps for constructing a high-precision virtual model of logistics equipment include: Preprocessing of equipment status data and environmental data outputs preprocessed model parameters. Preprocessing includes data normalization and missing value imputation. The normalization formula is:

[0048] Where d represents the original device status data and environmental data, d ’ The data are normalized, min(d) is the minimum data value, and max(d) is the maximum data value.

[0049] Based on the device state data and task requirements, a multi-layer modeling architecture is determined to obtain a virtual model. The multi-layer modeling architecture can adopt a hierarchical modeling method to divide the model into a physical layer, a behavior layer, and a function layer.

[0050] The pre-processed model parameters are input into the virtual model, and the model variables of the virtual model are calibrated through historical operation records to verify the consistency of the virtual model with the behavior of the logistics equipment. The calibration method can use the least squares method for calibration.

[0051] The subsystem models of mechanical motion, energy consumption, and sensor feedback are integrated to construct a high-precision dynamic interaction model. The dynamic interaction model can use a multi-body dynamics model.

[0052] A real-time data interaction channel between the virtual model and the high-precision dynamic interaction model and the sensor and control system is constructed to output a high-precision virtual model of the logistics equipment. The data interaction channel can use a message queue or a real-time database to ensure the real-time and accuracy of the data. The intelligent strategy optimization module also has a model updating function, which dynamically updates the virtual model according to real-time data feedback to adapt to changes in equipment state and environment.

[0053] The plurality of logistics equipment terminals are configured to receive and execute the control instructions and output execution logs.

[0054] The logistics equipment terminal includes an instruction parsing unit and an instruction execution unit. The instruction parsing unit is configured to disassemble the control instructions into underlying control parameters, and the instruction execution unit is configured to drive each device terminal according to the underlying control parameters and output execution logs. The step of disassembling the control instructions into underlying control parameters includes: Receiving control instructions from the upper system, parsing the instruction structure through ASCII / HEX protocol, extracting instruction header, operation code, device identifier, and target parameter field. ASCII / HEX protocol: a protocol for data communication, ASCII (American Standard Code for Information Interchange) is a standard for character encoding, and HEX (hexadecimal) is a base-16 counting system. This protocol is used to parse the structure of the control instructions and extract the instruction header, operation code, device identifier, and target parameter field.

[0055] In this process, the fault tolerance of protocol parsing needs to be ensured, such as retransmission request for instructions with checksum errors, and recording of abnormal logs. Checksum is a segment of additional data added during data transmission or storage to detect whether the data has been tampered with or damaged. By calculating and comparing the checksum, it can be determined whether the data has been tampered with or damaged during transmission or storage.

[0056] According to the operation code, a predefined instruction mapping table is matched to convert the high-level semantic instruction into a device layer action combination, and output the target parameter. The instruction mapping table is a predefined table used to convert high-level semantic instructions (i.e. more abstract instructions issued by users or upper systems) into device layer action combinations (i.e. specific operation steps that can be directly executed by devices). The operation code is an important index of the mapping table, and different operation codes correspond to different action sequences. Operation code 0x01 may represent "start the device", and the mapping table defines the specific action sequence required to start the device. It is necessary to support dynamic updating of the mapping table, avoid conflicts between new and old instructions through version control mechanism, and configure a bottom-up processing strategy for unknown operation codes.

[0057] The target parameter is split into multiple logical parameters in multiple dimensions, including numerical parameters, coordinate parameters, and time parameters. The numerical parameters are converted into PWM duty cycles through the PID formula, the coordinate parameters are converted into step motor steps through the kinematics model, and the time parameters are converted into timer interrupt periods through the RTC clock chip parameters. In this link, a parameter validity verification mechanism needs to be introduced, such as upper and lower limit constraints for numerical parameters, and anti-jitter filtering algorithm is used to eliminate sensor noise interference. The kinematics model is usually used to study the mathematical model of the geometric properties (such as position, velocity, acceleration, etc.) of the object motion and the relationship with time. In this case, it can be used to convert coordinate parameters into step motor steps, thereby controlling the precise motion of the step motor. The anti-jitter filtering algorithm is an algorithm used to eliminate sensor noise interference. Due to the interference of the external environment during the actual measurement of the sensor, the output signal may appear jitter. This algorithm processes the signal to remove these jitters, improving the stability and accuracy of the signal.

[0058] The multiple logical parameters are clustered and integrated to output the bottom layer control parameters. The clustering algorithm needs to be dynamically adjusted according to the topology relationship between devices, such as using a distributed clustering strategy based on the minimum spanning tree to reduce the computational complexity.

[0059] The fault feedback module is used to determine whether a fault occurs between multiple logistics devices according to the execution log, and output feedback information.

[0060] The fault feedback module includes a log analysis unit and a fault diagnosis unit. The log analysis unit is used to determine device coordination abnormalities based on a sliding window according to the execution log, identify the fault propagation path through association rule mining, and output feedback information.

[0061] The steps of identifying the fault propagation path through association rule mining include: A device adjacency matrix is constructed to define the workflow dependency between devices. The adjacency matrix needs to be dynamically updated, such as using an incremental updating algorithm to reduce the full reconstruction overhead, and incorporating the fault propagation probability attribute in the weight calculation.

[0062] Based on the workflow dependency and device collaboration anomaly, the key features of device operation anomaly are extracted. The key feature extraction needs to integrate time domain and frequency domain analysis methods, such as superimposing wavelet transform on vibration signals to extract multi-scale features.

[0063] According to the key features, the item set of association rule mining is defined, and the rule parameters of association rule mining are set. The rule parameters need to support adaptive adjustment, the feature importance is calculated by entropy method, and the rule confidence threshold is optimized based on Gini coefficient.

[0064] Using time series association rules, the rule parameters are combined to identify the fault propagation path. The identified results need to be verified for causality, such as using Granger causality test to eliminate pseudo-association.

[0065] Further, the fault diagnosis unit is used to establish an LSTM prediction model based on the historical fault knowledge base to provide early warning of faults according to feedback information. Lightweight edge computing nodes need to be deployed, and quantization compression models are used in time-sensitive scenarios to reduce delay, and multi-device collaborative evolution is realized through a federated learning framework. LSTM is a long short-term memory network, which is a special type of recurrent neural network (RNN) that can effectively solve the gradient vanishing and gradient explosion problems that occur when traditional RNNs process long sequence data. It introduces a gating mechanism that can selectively remember and forget information, thus better capturing long-range dependencies in time series data.

[0066] The steps of establishing an LSTM prediction model based on the historical fault knowledge base to provide early warning of faults include: Extract the fault features in the historical fault indication library, and perform feature statistics on the fault features according to the time window to output the time window features. Feature engineering needs to introduce an attention mechanism to adaptively select key features, and grid search optimization is performed on the time window parameters.

[0067] Divide the time window features into training set, validation set and test set in chronological order. Data division needs to consider the time series of time series, and uses an overlapping window strategy to avoid data leakage, and uses the SMOTE algorithm to enhance unbalanced data. SMOTE (Synthetic Minority Over-sampling Technique) algorithm is an algorithm for handling imbalanced data, which generates new samples by interpolating minority class samples to increase the number of minority class samples, thereby balancing the data set and improving the classification ability of the model for minority class samples.

[0068] The LSTM model is constructed and trained using the training set. The LSTM model loss is verified using the validation set, and when the model loss does not decrease for 5 consecutive rounds, the training is terminated. The model construction needs to integrate the attention mechanism layer to capture long-range dependencies and suppress overfitting through Dropout technology. During the training process, the risk of gradient explosion needs to be monitored in real time, and the gradient clipping threshold is set.

[0069] The confusion matrix, F1 score, ROC curve are calculated on the test set, and the early warning timeliness of the LSTM model for sudden failures is checked. The confusion matrix is a tool for evaluating the performance of a classification model, which shows the comparison between the predicted results and the actual results of different classes in the classification process. Through the confusion matrix, evaluation indicators such as accuracy, recall, and F1 score can be calculated. The F1 score is an evaluation indicator that considers both precision and recall, and is used to measure the performance of a classification model. It is the harmonic mean of precision and recall, with a value range of 0 to 1, and the larger the value, the better the model performance. The ROC curve is the Receiver Operating Characteristic Curve, which is used to show the performance of a classification model at different thresholds. The larger the area under the curve (AUC), the better the classification performance of the model. In the calculation process, additional MTTD and MTTR indicators need to be calculated, and the robustness of the test model is tested through adversarial sample generation. MTTD (Mean Time to Detect) is the average detection time, which refers to the average time required from the occurrence of a fault to its detection. This indicator is used to measure the timeliness of the fault warning system. MTTR (Mean Time to Repair) is the average repair time, which refers to the average time required from the detection of a fault to its repair. This indicator is used to measure the efficiency of fault repair.

[0070] The standby management module is used to determine whether the standby device needs to be enabled according to the equipment state data and feedback information. If needed, a standby scheduling strategy is output. The decision-making process needs to build a Markov decision process model, comprehensively consider the equipment utilization rate, maintenance cost and fault risk, and optimize the long-term benefit through reinforcement learning. Markov Decision Process (MDP) is a mathematical model used to describe the process of making decisions in an uncertain environment. It consists of state space, action space, state transition probability and reward function. By defining these elements, the optimal action can be selected in different states to maximize long-term benefits. Emergency solutions include cold standby mode (delayed start) and hot standby mode (seamless switching), and support pre-play mechanism based on digital twinning. Cold standby mode is a startup mode of standby device, when the main device fails, the standby device will not start immediately, but need a certain time for initialization and startup process, and then take over the work of the main device. Hot standby mode is a startup mode of standby device, standby device is in standby state at any time, when the main device fails, standby device can seamlessly switch to take over the work of the main device, to ensure the continuity and stability of the system. The pre-play mechanism of digital twinning is to simulate different fault scenarios and standby scheduling schemes to evaluate the feasibility and effectiveness of the scheme, and improve the reliability and efficiency of standby management.

[0071] In summary, the electric control system for logistics equipment provided by the application realizes the synchronous evolution of virtual model and physical equipment through the multi-layer high-precision dynamic interaction model constructed by the intelligent strategy optimization module, the history data calibration and real-time data interaction channel, the simulation test unit can predict the effect of strategy execution and dynamically adjust the weight, and the generated control instruction is processed through PID conversion, kinematics modeling and other multidimensional processing to ensure the accuracy and stability of the device action. The fault feedback module is based on sliding window statistics and association rule mining, which can accurately identify the fault propagation path, and through the LSTM prediction model combined with attention mechanism to realize fault warning. The deployment of lightweight edge computing node and federated learning framework supports low-latency response and multi-device collaborative evolution in time-sensitive scenarios, significantly reduces the average repair time and improves the system robustness. The standby management module evaluates the equipment state through Markov decision process, combines cold / hot standby mode and digital twinning pre-play mechanism, realizes seamless switching or delayed start of standby device. The design ensures the continuity of the system while optimizing the maintenance cost and fault risk, and the effectiveness of the emergency solution is further strengthened through simulation verification.

[0072] Those skilled in the art can clearly understand the technical solutions of the various embodiments from the above description of the embodiments, and the various embodiments can be implemented by means of software with the necessary general hardware platforms, and of course, can also be implemented by hardware. Based on such understanding, the above technical solutions, essentially or in other words, the part of the prior art that makes a contribution, can be embodied in the form of a software product, which can be stored in a computer readable storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a number of instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0073] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, rather than limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that: it can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacement for some technical features therein; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An electric control system for a logistics device, characterized in that, The application relates to a logistics equipment intelligent scheduling system, which comprises the following parts: a data acquisition module for acquiring equipment state data and environment data in real time; a main control module for generating a scheduling strategy according to the equipment state data and the environment data; an intelligent strategy optimization module for constructing a virtual model according to the equipment state data and the scheduling strategy and outputting a control instruction; a plurality of logistics equipment ends for receiving and executing the control instruction and outputting an execution log; a fault feedback module for judging whether faults occur among the plurality of logistics equipment according to the execution log and outputting feedback information; a standby line management module for judging whether standby equipment needs to be started according to the equipment state data and the feedback information; and outputting a standby line scheduling strategy if needed.

2. The electric control system for a logistics device according to claim 1, wherein The main control module comprises a task analysis unit and a strategy generation unit, the task analysis unit is used for formulating a task priority queue and a constraint condition set according to the equipment state data and the environment data, and the strategy generation unit is used for generating the scheduling strategy based on an operational research algorithm according to the task priority queue and the constraint condition set.

3. The electric control system for a logistics device according to claim 2, wherein The task analysis unit formulates a task priority queue and a constraint condition set in the following steps: data cleaning, feature extraction and standardization processing are performed on the equipment state data and environment data to construct a multi-dimensional data set containing equipment operation parameters, task space-time attributes and environment dynamic indexes; a multi-level evaluation model is used to calculate a task priority value according to the multi-dimensional data set to generate the task priority queue; a hierarchical constraint system is established according to the equipment operation parameters and the environment dynamic indexes to output an initial constraint condition set; a finite state machine model is used to detect conflicts of the task priority queue and the initial constraint condition set, and a constraint relaxation algorithm is used to dynamically adjust non-critical constraint parameters to output the constraint condition set.

4. The electric control system for a logistics device according to claim 1, wherein The intelligent strategy optimization module comprises a model construction unit, a simulation test unit and an instruction conversion unit; the model construction unit is used for constructing a high-precision virtual model of logistics equipment according to the equipment state data and the scheduling strategy; the simulation test unit is used for simulating the execution of the scheduling strategy through the high-precision virtual model of logistics equipment and the environment data to generate a simulation report; and the instruction conversion unit is used for adjusting a scheduling weight according to the simulation report and generating the control instruction according to the scheduling weight.

5. The electric control system for a logistics device according to claim 4, wherein The model construction unit constructs the high-precision virtual model of logistics equipment in the following steps: preprocessing is performed on the equipment state data and the environment data to output preprocessing model parameters; a multi-layer modeling architecture is determined based on the equipment state data and task requirements to obtain a virtual model; the preprocessing model parameters are input into the virtual model, model variables of the virtual model are calibrated through historical operation records, and the virtual model is verified to be consistent with behaviors of the logistics equipment; subsystem models of mechanical movement, energy consumption and sensor feedback are integrated to construct a high-precision dynamic interaction model; real-time data interaction channels of the virtual model, the high-precision dynamic interaction model, the sensor and the control system are constructed to output the high-precision virtual model of logistics equipment.

6. The electric control system for a logistics device according to claim 1, wherein The logistics equipment end comprises an instruction analysis unit and an instruction execution unit, the instruction analysis unit is used for disassembling the control instruction into bottom layer control parameters, and the instruction execution unit is used for driving each equipment terminal according to the bottom layer control parameters and outputting an execution log.

7. An electric control system for a logistics device according to claim 6, characterized in that, The step of disassembling the control instruction into bottom layer control parameters by the instruction analysis unit comprises: receiving a control instruction from an upper layer system, analyzing an instruction structure through an ASCII / HEX protocol, extracting an instruction header, an operation code, a device identifier and a target parameter field; according to the operation code, matching a pre-defined instruction mapping table, converting a high layer semantic instruction into a device layer action combination, and outputting a target parameter; performing multi-dimensional splitting of the target parameter into a plurality of logical parameters, including a numerical value type parameter, a coordinate parameter and a time parameter; converting the numerical value type parameter into a PWM duty ratio through a PID formula, converting the coordinate parameter into a step motor step number through a kinematics model, and converting the time parameter into a timer interrupt period through an RTC clock chip parameter; clustering and integrating a plurality of the logical parameters to output the bottom layer control parameters.

8. The electric control system for a logistics device according to claim 1, wherein The fault feedback module comprises a log analysis unit and a fault diagnosis unit, the log analysis unit is used for, according to the execution log, statistically analyzing device cooperation abnormalities based on a sliding window, identifying a fault propagation path through association rule mining, and outputting the feedback information; the fault diagnosis unit is used for, according to the feedback information, establishing an LSTM prediction model based on a historical fault knowledge base to give a fault warning.

9. The electric control system for a logistics device according to claim 8, wherein The step of the log analysis unit identifying a fault propagation path through association rule mining comprises: constructing a device adjacency matrix to define a work flow dependence between devices; based on the work flow dependence and the device cooperation abnormalities, extracting key features of device operation abnormalities; according to the key features, defining an item set of association rule mining, and setting rule parameters of the association rule mining; using a time sequence association rule, combining the rule parameters to identify the fault propagation path.

10. The electric control system for a logistics device according to claim 8, wherein The step of the fault diagnosis unit establishing an LSTM prediction model based on a historical fault knowledge base to give a fault warning comprises: extracting fault features in the historical fault indication library, and performing feature statistics on the fault features according to a time window to output time window features; dividing the time window features into a training set, a verification set and a test set in chronological order; performing LSTM model construction, and using the training set to train the LSTM model; using the verification set to verify a loss of the LSTM model, and when the model loss does not decrease continuously for multiple rounds, terminating the training; checking the warning timeliness of the LSTM model on the test set for a sudden fault.

Images