A transport vehicle safety control system and method

Abstract

The present invention belongs to the technical field of transportation vehicle safety management and control, and in particular, relates to a transportation vehicle safety management and control system and method. The system integrates four modules: a central control platform, vehicle access, transportation safety monitoring, and vehicle health management. The central control platform is the core, responsible for system deployment, operation, and instruction issuance. The vehicle access module ensures the identity and qualification review of transportation vehicles and guarantees regional entry and exit safety management. The transportation safety monitoring module monitors the status of transportation vehicles and drivers in real time, warns of safety hazards through intelligent analysis and responds quickly. The vehicle health management module regularly monitors the status of transportation vehicles, predicts faults, and implements preventive maintenance to extend the life of transportation vehicles. The present invention uses intelligent means to achieve closed-loop control from access to monitoring to health management, which significantly reduces transportation risks and improves operational reliability.

Description

Technical Field

[0001] The present invention belongs to the technical field of transportation vehicle safety management and control, and in particular, relates to a transportation vehicle safety management and control system and method. Background Art

[0002] With the rapid development of the logistics industry, transport vehicles, as a key link in the logistics chain, have become the focus of industry attention in terms of safety management. Traditional manual management methods can no longer meet the modern transportation industry's demand for high efficiency and high safety. On the one hand, the management of vehicle access and task execution relies on manual review and scheduling, which is not only time-consuming and labor-intensive, but also prone to human errors, leading to safety hazards. On the other hand, there is insufficient real-time monitoring during transportation, and the ability to detect abnormalities in vehicle operating status and driver behavior is limited, making it difficult to detect and deal with potential risks in a timely manner. This not only affects transportation efficiency, but also increases the possibility of accidents.

[0003] For example, the patent with publication number CN118096000A discloses a hazardous chemical vehicle transportation safety management and control system and method, including: a driver identity information recognition device, a vehicle identity information recognition device, a task confirmation device, a vehicle autonomous positioning module, a monitoring device, an AI-based intelligent analysis platform, a checkpoint control module, and a vehicle operation control module.

[0004] For example, the patent with publication number CN111369185A discloses a method, system, equipment and storage medium for overweight control of transport vehicles. The method includes: respectively obtaining the marked weight and actual load of the transport vehicle; comparing the obtained marked weight and actual load of the transport vehicle to determine whether the transport vehicle is overloaded; if so, sending the overload information of the transport vehicle to the big data platform, and obtaining the location information of the transport vehicle at the same time; sending the received overload information and location information of the transport vehicle to nearby traffic management checkpoints through the big data platform to remind management personnel to intercept and punish the transport vehicle.

[0005] The above existing technologies all have the following problems: 1) The existing systems are often too rigid in controlling the entry and exit of vehicles and cannot be flexibly adjusted according to the actual mission status of the vehicle, resulting in waste of resources or security loopholes; 2) In the face of abnormal situations occurring during transportation, the system warning is not timely and the emergency response processing strategy lacks pertinence; In order to solve the above problems, the present invention provides a transportation vehicle safety management and control system and method. Summary of the invention

[0006] In view of the shortcomings of the prior art, the present invention proposes a transport vehicle safety management and control system and method. The system integrates four major modules: a central control platform, vehicle access, transport safety monitoring and vehicle health management. The central control platform is the core, responsible for system deployment, operation and command issuance. The vehicle access module ensures the identity and qualification review of the transport vehicle and guarantees the safety control of regional entry and exit. The transport safety monitoring module monitors the status of the transport vehicle and the driver in real time, warns of safety hazards through intelligent analysis and responds quickly. The vehicle health management module regularly monitors the status of the transport vehicle, predicts faults and implements preventive maintenance to extend the life of the transport vehicle. The present invention realizes closed-loop control from access to monitoring and then to health management through intelligent means, which significantly reduces transportation risks and improves operational reliability.

[0007] To achieve the above object, the present invention provides the following technical solutions:

[0008] A transport vehicle safety management and control system, comprising: a central control platform, a vehicle access module and a transport safety monitoring module;

[0009] The central control platform is used for the deployment and operation of all modules and units in the security management and control system, as well as the uploading of operation results and the allocation and issuance of execution task instructions;

[0010] The vehicle access module includes a vehicle information database unit, a tag identification unit and an access control unit;

[0011] The vehicle information database unit is used to construct a transport vehicle information database using a natural language algorithm based on the acquired transport vehicle attributes and qualification text information, and configure a qualification review model at the transport vehicle information database interface position to conduct qualification review on the transport vehicle information added in real time, and complete pre-registration and transport vehicle information storage for the transport vehicles that have passed the review; the label recognition unit is used to match and identify the transport vehicle label information and the pre-registered information in the transport vehicle information database based on the electronic label installed on the transport vehicle through a matching recognition algorithm to confirm the identity of the transport vehicle; the entry and exit control unit is used to construct entry and exit discrimination rules, and automatically determine whether the transport vehicle is allowed to enter and exit the specified area by constructing entry and exit discrimination rules based on the matching and recognition results of the label recognition unit and the task instructions issued by the central control platform;

[0012] The transportation safety monitoring module includes a transportation vehicle monitoring unit, an abnormal trajectory recognition unit, and a behavior recognition unit;

[0013] The vehicle monitoring unit is used to monitor and obtain the operation attribute data, position and trajectory data of the transport vehicles that are performing tasks in the transport vehicle information database in real time through the deployed integrated sensors, as well as the video and image data of the driver during driving; the abnormal trajectory recognition unit is used to monitor the abnormality of the transport vehicle's driving trajectory and the transport vehicle's operation status score in real time through the trajectory recognition algorithm and the comprehensive evaluation algorithm based on the real-time operation trajectory and position data of the transport vehicle and the operation attribute data of the transport vehicle; the behavior recognition unit is used to monitor the driver's micro-expressions and abnormal behavior status during driving in real time through the behavior and micro-expression recognition model based on the video and image data of the driver during driving;

[0014] The designated area includes the parking area and the working area. The specific implementation steps for entering and exiting the control unit include:

[0015] B1. When transport vehicle A is not performing a task, the label information of transport vehicle A is identified by the label identification unit to obtain the corresponding label information of transport vehicle A, and the matching algorithm is used to match the obtained label information of transport vehicle A with the pre-registered information in the transport vehicle information database. , the specific formula is:

[0016] ,

[0017] in, , and It represents the similarity between the license plate number, vehicle model and driver information corresponding to the label information obtained by identifying the transport vehicle A and the pre-registered information in the transport vehicle information database, It indicates identifying the transport vehicle A and obtaining the corresponding license plate number information in the label information. Indicates the pre-registered license plate number information in the transport vehicle information database. It indicates identifying the transport vehicle A and obtaining the corresponding vehicle model information in the label information. Indicates the vehicle model information pre-registered in the transport vehicle information database. It indicates identifying the transport vehicle A and obtaining the corresponding driver information in the label information. Represents the driver information pre-registered in the transport vehicle information database; represents the corresponding weighting coefficient;

[0018] B2. Set the tag information matching threshold , the obtained matching degree and For comparison, When , transport vehicle A is allowed to enter and exit the parking area, but is prohibited from entering and exiting the work area. When the vehicle is in the parking area, only transport vehicle A is allowed to leave the parking area, and is prohibited from entering the parking area and entering and exiting the work area;

[0019] B3: When transport vehicle A performs a task, the label information of transport vehicle A is judged through B1 and B2. If it satisfies If the condition is not met, the task instruction of transport vehicle A is judged again. Then they are directly prohibited from entering the work area;

[0020] The specific steps of secondary identification of the task instruction of transport vehicle A include:

[0021] B4, obtain the task instruction information of transport vehicle A through sensors, and use visual sensors and weight sensors to obtain the type and weight information of the goods transported by transport vehicle A at the current moment;

[0022] B5. Based on the task instruction information obtained in B4 and the type and weight information of the goods transported by vehicle A at the current moment, a task instruction verification process formula is constructed through Boolean functions. The task instruction information of transport vehicle A is compared with the information in the preset task list of the central control platform. If all the comparisons and judgments are passed, transport vehicle A is allowed to enter and exit the working area and parking area. The specific formula of the task instruction verification process is:

[0023] ,

[0024] in, It means to verify whether the task instruction of transport vehicle A exists in the preset task list. It verifies whether the task execution time of transport vehicle A is within the allowed time window. It verifies whether the mission purpose in the mission instruction of transport vehicle A matches the current destination. It verifies whether the authority level of transport vehicle A matches the task requirements. It verifies whether the cargo type currently transported by transport vehicle A matches the cargo type in the task instruction. It verifies whether the current cargo weight of transport vehicle A is consistent with the cargo weight in the task instruction. Represents the "and" operation in the logical operator;

[0025] B6. If one of the comparisons fails, vehicle A is prohibited from entering or exiting the work area, but is allowed to enter or exit the parking area;

[0026] The specific steps of constructing the entry and exit discrimination rules in the entry and exit control unit include:

[0027] According to the B1-B6 process, the entry and exit discrimination rules are constructed, and the constructed entry and exit discrimination rules are built into the entry and exit control unit to discriminate the transport vehicles entering and leaving the specified area. The specific formula of the entry and exit discrimination rules is:

[0028] ,

[0029] Where A=S&A W means that transport vehicle A is allowed to enter and exit the parking area, but is prohibited from entering and exiting the working area; A S&A W means that transport vehicle A is prohibited from entering or leaving the parking area and the working area. Represents the process formula verified by task instructions Carry out comparison and identification, and all the identifications pass. Represents the process formula verified by task instructions Perform a comparison and judgment to determine if at least one of the verifications failed, A=S&A W means that the transport vehicle A is allowed to enter and exit the parking area and the working area; A means the transport vehicle A, S means the parking area, W means the working area, T means that v is judged to be true, that is, True, and F means that v is judged to be false, that is, v is False;

[0030] The task instructions issued by the central control platform include the starting point, transportation time, destination and optimal driving path of the transport vehicle. The optimal driving path is automatically generated by the central control platform according to the task instruction information issued and the road condition information obtained in real time through the path planning generation algorithm, and transmitted to the corresponding transport vehicle; the specific steps of the abnormal trajectory recognition unit monitoring the abnormal driving trajectory of the transport vehicle through the trajectory recognition algorithm include:

[0031] C1. Set the trajectory deviation threshold , and monitor the process trajectory data of the transport vehicle traveling along the optimal driving path in real time. According to the coordinate information of each position in the process trajectory data and the position information in the optimal driving trajectory in the task instruction, the position difference is calculated through the spatial distance formula. ,in It represents the distance between the i-th position coordinate in the real-time monitoring process trajectory data and the corresponding position coordinate in the optimal driving trajectory;

[0032] C2, the position difference obtained Threshold for deviation from track For comparison, when N consecutive position differences in the real-time monitoring process trajectory data are greater than the trajectory deviation threshold When , it is determined that the vehicle position is offset;

[0033] C3. When the position is offset, the current position coordinates after the offset, the current road condition information and the target point position are monitored, and the path planning algorithm is used to determine whether the current position coordinates after the offset can reach the target point position within the transportation time length;

[0034] C4. If it is determined that the transport vehicle cannot reach the target point or cannot reach the target point within the transport time, the trajectory deviation is determined to be abnormal, and a trajectory deviation abnormality warning is sent to the vehicle;

[0035] C5. When it is determined that the transport vehicle can reach the target point within the transport time, a new optimal driving path trajectory is generated based on the current position coordinates and the target point position after the offset, and sent to the corresponding transport vehicle. The corresponding transport vehicle is continuously monitored on the new optimal driving path trajectory to determine whether a secondary trajectory offset abnormality warning occurs.

[0036] Specifically, the specific workflow of the vehicle information library unit includes:

[0037] A1. Build an integrated model of image text extraction and entity-relationship joint extraction based on the CTPN pre-trained sub-model, the Chinese pre-trained BERT sub-model, the BILSTM sub-model and the CRF sub-model, obtain vehicle information text data and driver information data, and use the obtained vehicle information text data to fine-tune the integrated model of image text extraction and entity-relationship joint extraction to obtain the trained integrated model of image text extraction and entity-relationship joint extraction;

[0038] A2. The trained integrated model of image text extraction and entity-relationship joint extraction is built into the Neo4j graph database. When the driver uploads the corresponding text information, the built-in integrated model of image text extraction and entity-relationship joint extraction is used to perform image or text recognition and entity relationship extraction on the information uploaded by the driver to obtain vehicle information and driver entity relationship data;

[0039] A3. Build and train the qualification review model based on the Chinese pre-trained BERT sub-model and the cross-attention sub-model, embed the obtained qualification review model into the Neo4j graph database interface, match the vehicle information and driver entity relationship data obtained in A2 with the qualification review conditions configured in the Neo4j graph database, and obtain the matching similarity ;

[0040] A4. Set the audit match threshold. When the matching similarity is greater than or equal to the audit match threshold, the audit is automatically passed, and the acquired vehicle information and driver entity relationship data are automatically saved in the transport vehicle information database built by the Neo4j graph database to complete the vehicle information storage and driver information pre-registration.

[0041] Specifically, the specific steps of the behavior recognition unit monitoring the driver's micro-expressions and abnormal behavior during driving include:

[0042] D1. Obtain historical driver behavior and micro-expression abnormal video and image data during driving, and perform frame processing on the acquired video data through the OpenCV library to obtain abnormal micro-expression and behavior image data, and obtain the corresponding audio data when abnormal behavior and micro-expression occur from the acquired behavior and micro-expression abnormal video data through audio extraction technology;

[0043] D2. Annotate the above-obtained abnormal behavior and micro-expression video, image and audio data with corresponding abnormal labels, and annotate the obtained abnormal behavior and micro-expression video and image data with abnormal behavior types and scores. At the same time, annotate the corresponding abnormal types for the audio data, and annotate the audio data with scores based on the intensity of the abnormal sounds in the audio data;

[0044] D3. Construct a behavior and micro-expression recognition model, input the acquired original abnormal behavior and micro-expression video data into the video coding layer of the behavior and micro-expression recognition model, output the abnormal behavior and micro-expression time change features, input the annotated behavior and micro-expression abnormal image data into the image coding layer, output the abnormal behavior and micro-expression static features, input the annotated audio data into the audio coding layer, and output the corresponding abnormal behavior audio features;

[0045] D4. Input the acquired abnormal behavior and micro-expression time change features, abnormal behavior and micro-expression static features, and abnormal behavior audio features into the attention fusion layer constructed by the collaborative attention network for fusion, and obtain abnormal behavior and micro-expression fusion features;

[0046] D5. Input the acquired abnormal behavior and micro-expression fusion features into output layer 1 and output layer 2 at the same time, and obtain the abnormal behavior type and the corresponding score respectively. and abnormal micro-expression types and corresponding scores ;

[0047] D6. According to the corresponding abnormal behaviors and micro-expression types and corresponding scores, set the cross entropy loss function and the corresponding loss threshold to train the D1-D5 process to obtain the trained behavior and micro-expression recognition model;

[0048] Specifically, the specific steps of the behavior recognition unit monitoring the driver's micro-expressions and abnormal behavior during driving include:

[0049] D7. The trained behavior and micro-expression recognition model is built into the behavior recognition unit. During the driving of the transport vehicle, the driver's behavior, micro-expression and audio data are obtained in real time through the configured visual sensor;

[0050] D8. Input the acquired driving behavior, micro-expression and audio data into the behavior recognition unit to obtain the corresponding abnormal behavior and micro-expression type and the corresponding score and ;

[0051] D9. Set abnormal behavior score threshold and abnormal micro-expression score threshold ,when and When the driver drives normally, or It determines when the driver's driving is abnormal, and feeds back the acquired abnormal behavior and micro-expression types to the central control platform, generates corresponding processing strategies, and transmits them to the corresponding vehicle drivers in real time, and issues early warnings for abnormalities passing through the on-board sensors of transport vehicles.

[0052] A method for safety management and control of transport vehicles, the specific steps comprising:

[0053] S1. Obtain the attributes and qualification text information of transport vehicles, build a transport vehicle information database through natural language algorithms, and configure a qualification review model to review the transport vehicle information added in real time;

[0054] S2. According to the acquired transport vehicle information database, the transport vehicle is identified through the electronic tag and matching recognition algorithm. According to the confirmation result and the task instruction issued by the central control platform, the entry and exit discrimination rules are used to determine whether the transport vehicle can enter and exit the specified area, and the task is executed;

[0055] S3. When it is determined that the transport vehicle is capable of performing the task, the operating attribute data, position and trajectory data of the transport vehicle performing the task and the video and image data of the driver during driving are simultaneously acquired through real-time monitoring by sensors;

[0056] S4. Based on the operation attribute data, location and trajectory data of the transport vehicle, the abnormal driving trajectory of the transport vehicle and the operation status score of the transport vehicle are obtained through the trajectory recognition algorithm and the comprehensive evaluation algorithm. At the same time, based on the video and image data of the driver during driving, the micro-expression and abnormal behavior status of the driver during driving are obtained through the behavior and micro-expression recognition model;

[0057] S5. Based on the acquired abnormal driving trajectory and operating status scores of the transport vehicle and the driver's micro-expressions and abnormal behavior during driving, the warning and response measurement and processing library built into the central control platform is used to issue warnings for the identified abnormal types through the on-board equipment. At the same time, based on the corresponding abnormal type results, the corresponding emergency response processing strategy is generated and issued to the corresponding abnormal transport vehicle through matching and generation algorithms.

[0058] A computer-readable storage medium stores computer instructions, which, when executed, execute a method for safety management and control of transport vehicles.

[0059] Compared with the prior art, the present invention has the following beneficial effects:

[0060] 1. Aiming at the deficiencies of the prior art, the present invention uses an intelligent vehicle access module to enable the system to efficiently conduct qualification review and dynamic management of transport vehicles, avoiding rigid access control, ensuring that only vehicles that meet the standards can perform tasks, and effectively preventing waste of resources and potential security loopholes. In addition, the system can flexibly adjust access control according to the actual task status of the vehicle, such as through accurate label recognition and secondary identification of task instructions, to achieve intelligent management of transport vehicles in parking areas and working areas, enhancing the adaptability and safety of the system.

[0061] 2. In view of the deficiencies in the prior art, the abnormal trajectory recognition and behavior recognition functions in the transportation safety monitoring module of the present invention can monitor the abnormal driving trajectory of the transport vehicle and the driver's micro-expressions and behavioral status in real time. Once deviation from the preset route or abnormal behavior of the driver is detected, the system can respond quickly, issue timely warnings and take targeted measures, which greatly improves the emergency response capability during transportation and reduces the risk of accidents. In particular, by constructing a refined behavior and micro-expression recognition model, the system can accurately evaluate the driver's status and ensure the safety of transportation operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] Figure 1 This is a module diagram of a transport vehicle safety management and control system according to Embodiment 1 of the present invention;

[0063] Figure 2 This is a structural diagram of the behavior and micro-expression recognition model of Example 1 of the present invention;

[0064] Figure 3 This is a schematic diagram of the working process principle of a transport vehicle safety management and control system according to Embodiment 2 of the present invention;

[0065] Figure 4 This is a flow chart of a method for safety management and control of transport vehicles according to Example 3 of the present invention. DETAILED DESCRIPTION

[0066] Example 1

[0067] See also Figure 1 , an embodiment provided by the present invention: a transport vehicle safety management and control system, comprising: a central control platform, a vehicle access module, a transport safety monitoring module and a vehicle health management module;

[0068] The central control platform is used for the deployment and operation of all modules and units in the security management and control system, as well as the uploading of operation results and the allocation and issuance of execution task instructions;

[0069] The vehicle access module is used to identify the identity information of the transport vehicle, review the qualification information and determine the access information of the specific area; the vehicle access module includes a vehicle information library unit, a label recognition unit and an access control unit;

[0070] The vehicle information database unit is used to construct a transport vehicle information database using a natural language algorithm based on the acquired transport vehicle attributes and qualification text information, and configure a qualification review model at the transport vehicle information database interface position to conduct qualification review on the transport vehicle information added in real time, and complete pre-registration and transport vehicle information storage for the transport vehicles that have passed the review; further, the specific steps of constructing the transport vehicle information database in this embodiment include:

[0071] A1. Based on the CTPN pre-trained sub-model, the Chinese pre-trained BERT sub-model, the BILSTM sub-model and the CRF sub-model, an integrated model of image text extraction-entity-relationship joint extraction is constructed to obtain vehicle information text data and driver information data, and the obtained vehicle information text data is used to fine-tune the integrated model of image text extraction-entity-relationship joint extraction to obtain a trained integrated model of image text extraction-entity-relationship joint extraction; further, in this embodiment, the vehicle information text data includes image and text information data such as vehicle registration documents, operating licenses, insurance certificates, driver's licenses and vehicle registration certificates, and the driver information data includes the driver's health data and age data, etc.;

[0072] A2. The trained integrated model of image text extraction and entity-relationship joint extraction is built into the Neo4j graph database. When the driver uploads the corresponding text information, the built-in integrated model of image text extraction and entity-relationship joint extraction is used to perform image or text recognition and entity relationship extraction on the information uploaded by the driver to obtain vehicle information and driver entity relationship data;

[0073] A3. Build and train the qualification review model based on the Chinese pre-trained BERT sub-model and the cross-attention sub-model, embed the obtained qualification review model into the Neo4j graph database interface, match the vehicle information and driver entity relationship data obtained in A2 with the qualification review conditions configured in the Neo4j graph database, and obtain the matching similarity ; Further, the qualification review conditions in this embodiment are set by those skilled in the art when constructing a transport vehicle information database, according to specific review requirements, such as age below 40, good health, etc.;

[0074] A4. Set the audit matching threshold. When the matching similarity If it is greater than or equal to the audit match threshold, it will automatically pass the audit, and the acquired vehicle information and driver entity relationship data will be automatically saved in the transportation vehicle information database built by the Neo4j graph database to complete the vehicle information storage and driver information pre-registration. Furthermore, in this embodiment, the audit match threshold is 0.9;

[0075] The tag identification unit is used to match and identify the tag information of the transport vehicle and the pre-registered information in the transport vehicle information database through a matching identification algorithm according to the electronic tag installed on the transport vehicle, so as to confirm the identity of the transport vehicle; further, in this embodiment, the tag identification unit captures the signal emitted from the electronic tag on the transport vehicle through a radio frequency identification (RFID) reader or a similar sensor; further, the matching identification algorithm in this embodiment realizes matching and identification of the corresponding vehicle information through a qualification review model;

[0076] The entry and exit control unit is used to automatically determine whether the transport vehicle is allowed to enter and exit the specified area according to the matching recognition result of the tag recognition unit and the task instruction issued by the central control platform through the entry and exit discrimination rules; further, in this embodiment, the specified area includes a parking area and a working area, and the parking area and the working area are circled in the electronic map through the electronic fence technology, and the circled and marked parking area and the working area are transmitted to the control system of the pre-registered transport vehicle in the transport vehicle information database through the central control platform;

[0077] The specific implementation steps of the access control unit include:

[0078] B1. When transport vehicle A is not performing a task, the label information of transport vehicle A is identified by the label identification unit to obtain the corresponding label information of transport vehicle A, and the matching algorithm is used to match the obtained label information of transport vehicle A with the pre-registered information in the transport vehicle information database. , the specific formula is:

[0079] ,

[0080] in, , and It represents the similarity between the license plate number, vehicle model and driver information corresponding to the label information obtained by identifying the transport vehicle A and the pre-registered information in the transport vehicle information database, It indicates identifying the transport vehicle A and obtaining the corresponding license plate number information in the label information. Indicates the pre-registered license plate number information in the transport vehicle information database. It indicates identifying the transport vehicle A and obtaining the corresponding vehicle model information in the label information. Indicates the vehicle model information pre-registered in the transport vehicle information database. It indicates identifying the transport vehicle A and obtaining the corresponding driver information in the label information. Represents the driver information pre-registered in the transport vehicle information database; represents the corresponding weighting coefficient;

[0081] B2. Set the tag information matching threshold , the obtained matching degree and For comparison, When , transport vehicle A is allowed to enter and exit the parking area, but is prohibited from entering and exiting the work area. When the vehicle is in the parking area, only transport vehicle A is allowed to leave the parking area, and is prohibited from entering the parking area and entering and exiting the work area;

[0082] B3: When transport vehicle A performs a task, the label information of transport vehicle A is judged through B1 and B2. If it satisfies If the condition is not met, the task instruction of transport vehicle A is judged again. Entering the work area is prohibited.

[0083] Specifically, the specific steps of performing secondary determination on the task instruction of the transport vehicle A include:

[0084] B4, obtain the task instruction information of transport vehicle A through sensors, and use visual sensors and weight sensors to obtain the type and weight information of the goods transported by transport vehicle A at the current moment;

[0085] B5. Based on the task instruction information obtained in B4 and the type and weight information of the goods transported by vehicle A at the current moment, a task instruction verification process formula is constructed through Boolean functions. The task instruction information of transport vehicle A is compared with the information in the preset task list of the central control platform. If all the comparisons and judgments are passed, transport vehicle A is allowed to enter and exit the working area and parking area. The specific formula of the task instruction verification process is:

[0086] ,

[0087] in, It verifies whether the task instruction of transport vehicle A exists in the preset task list, that is, checks whether the corresponding task ID and task description match the current task; It verifies whether the task execution time of transport vehicle A is within the allowed time window to prevent expired tasks or future tasks from being executed; It verifies whether the mission purpose in the mission instruction of transport vehicle A matches the current destination to ensure that it does not enter the unauthorized area by mistake; It means verifying whether the authority level of transport vehicle A matches the task requirements, that is, whether the current transport vehicle A meets the corresponding rated cargo capacity in the current task requirements; It verifies whether the current cargo type of transport vehicle A matches the cargo type in the task instruction, ensuring that the transported cargo matches the task instruction; It verifies whether the current cargo weight of transport vehicle A is consistent with the cargo weight in the task instruction. Represents the "and" operation in the logical operator;

[0088] B6. If one of the comparisons fails, vehicle A is prohibited from entering or exiting the work area, but is allowed to enter or exit the parking area;

[0089] B7. Construct an entry and exit discrimination rule according to the process of B1-B6, and embed the constructed entry and exit discrimination rule into the entry and exit control unit to discriminate the transport vehicles entering and exiting the specified area. The specific formula of the entry and exit discrimination rule is:

[0090] ,

[0091] Where A=S&A W means that transport vehicle A is allowed to enter and exit the parking area, but is prohibited from entering and exiting the working area; A S&A W means that transport vehicle A is prohibited from entering or leaving the parking area and the working area. Represents the process formula verified by task instructions Carry out comparison and judgment, and all the judgments pass. Represents the process formula verified by task instructions Perform a comparison and judgment to determine if at least one of the verifications failed, A=S&A W means that the transport vehicle A is allowed to enter and exit the parking area and the working area; A represents the transport vehicle A, S represents the parking area, W represents the working area, T represents that v is judged to be true, that is, True, and F represents that v is judged to be false, that is, v is False.

[0092] Through the above detailed access control process, the system realizes the refined management of transport vehicles, ensuring that only legally registered and current mission-compliant vehicles can enter the work area, effectively preventing unauthorized vehicles from entering, reducing safety hazards, and improving logistics efficiency and safety. At the same time, through intelligent access identification rules, human operation errors are reduced, the reliability and automation level of the system are enhanced, and a strong technical support is provided for modern logistics management. This system is particularly suitable for large logistics parks, warehouses, factories and other places. By accurately controlling the access of vehicles, the efficiency and safety of logistics operations are greatly improved. At the same time, it also provides managers with real-time vehicle status and location information for scheduling and monitoring.

[0093] The transportation safety monitoring module is used to monitor the driving status of the transportation vehicle and the driver's behavior status in real time, and to identify and warn of potential safety hazards during the operation of the transportation vehicle; the transportation safety monitoring module includes a vehicle monitoring unit, a storage unit, an abnormal trajectory recognition unit, a behavior recognition unit, and a warning and response unit;

[0094] The vehicle monitoring unit is used to monitor and obtain the operation attribute data, location and trajectory data of the transport vehicles on the mission in the transport vehicle information database and the video and image data of the driver during driving in real time through the deployed integrated sensors;

[0095] The storage unit is used for preprocessing, uniformly formatting and integrating the data acquired by the transport vehicle monitoring unit, and storing the integrated monitoring data in a distributed database. At the same time, the distributed database is configured with data update and deletion functions to automatically update or delete the data;

[0096] The abnormal trajectory recognition unit is used to monitor the abnormal driving trajectory of the transport vehicle and the running status score of the transport vehicle in real time through the trajectory recognition algorithm and the comprehensive evaluation algorithm according to the real-time position and trajectory data of the transport vehicle and the running attribute data of the transport vehicle; further, in this embodiment, the task instruction issued by the central control platform includes the starting point, transportation time length, target point and optimal driving path trajectory of the transport vehicle. The optimal driving path trajectory is automatically generated by the central control platform according to the task instruction information issued and the road condition information obtained in real time through the path planning generation algorithm, and transmitted to the corresponding transport vehicle; the specific steps of the abnormal trajectory recognition unit monitoring the abnormal driving trajectory of the transport vehicle through the trajectory recognition algorithm include:

[0097] C1. Set the trajectory deviation threshold , and monitor the process trajectory data of the transport vehicle traveling along the optimal driving path in real time. According to the coordinate information of each position in the process trajectory data and the position information in the optimal driving trajectory in the task instruction, the position difference is calculated through the spatial distance formula. ,in It represents the distance between the i-th position coordinate in the real-time monitoring process trajectory data and the corresponding position coordinate in the optimal driving trajectory;

[0098] C2, the position difference obtained Threshold for deviation from track For comparison, when N consecutive position differences in the real-time monitoring process trajectory data are greater than the trajectory deviation threshold When , it is determined that the vehicle position is offset;

[0099] C3. When the position is offset, the current position coordinates after the offset, the current road condition information and the target point position are monitored, and the path planning algorithm is used to determine whether the current position coordinates after the offset can reach the target point position within the transportation time length;

[0100] C4. If it is determined that the transport vehicle cannot reach the target location or cannot reach the target location within the transport time, the trajectory deviation is determined to be abnormal, and a trajectory deviation abnormality warning is sent to the vehicle;

[0101] C5. When it is determined that the transport vehicle can reach the target point within the transport time, a new optimal driving path trajectory is generated based on the current position coordinates and the target point position after the current offset, and sent to the corresponding transport vehicle. The corresponding transport vehicle is continuously monitored on the new optimal driving path trajectory to determine whether a secondary trajectory offset abnormality warning occurs.

[0102] Furthermore, in this embodiment, the specific steps of real-time monitoring of the transport vehicle operation status score through the comprehensive evaluation algorithm include:

[0103] C6. Obtaining the operation attribute data of the historical driving process of the transport vehicle and preprocessing the acquired data;

[0104] C7. Use the comprehensive fuzzy algorithm to build a comprehensive evaluation model for transport vehicles, input the pre-processed historical driving process operation attribute data into the comprehensive evaluation model for pre-training, and build the pre-trained comprehensive evaluation model into the on-board integrated sensor of the transport vehicle;

[0105] C8. Use integrated sensors to monitor and obtain the operation attribute data of transport vehicles in real time, and use the built-in comprehensive evaluation model to conduct real-time comprehensive scoring monitoring of the vehicle driving status based on the acquired operation attribute data of transport vehicles to obtain the operation status score of the transport vehicles;

[0106] C9. Set a threshold for the transport vehicle operation status score. When the transport vehicle operation status score obtained through real-time monitoring is greater than the transport vehicle operation status score threshold, the vehicle is judged to be driving abnormally, and an early warning is issued through the on-board early warning device. The obtained transport vehicle operation status score is uploaded to the central control platform, and fed back to the early warning and response unit to generate a corresponding emergency response processing strategy, which is transmitted to the transport vehicle judged to be abnormal.

[0107] The behavior recognition unit is used to monitor the driver's micro-expressions and abnormal behavior status during driving in real time through the behavior and micro-expression recognition model based on the video and image data obtained during the driver's driving process; for further information, please refer to Figure 2 In this embodiment, the specific steps of the behavior recognition unit monitoring the driver's micro-expressions and abnormal behavior during driving include:

[0108] D1. Obtain the historical driver's behavior and micro-expression abnormal video and image data during driving, and perform frame processing on the acquired abnormal video data through the OpenCV library to obtain abnormal micro-expression and behavior image data, and obtain the corresponding audio data when the abnormal behavior and micro-expression occur from the acquired behavior and micro-expression abnormal video data through audio extraction technology;

[0109] D2. Annotate the above-obtained abnormal behavior and micro-expression video and image data, as well as the corresponding audio data when the abnormal behavior and micro-expression occur, with corresponding abnormal labels, and annotate the acquired abnormal behavior and micro-expression video and image data with abnormal behavior types and scores, and annotate the audio data with corresponding abnormal types, and annotate the audio data with scores based on the intensity of abnormal sounds in the audio data;

[0110] Furthermore, the video data annotation in this embodiment includes abnormal timestamp annotation, specifically: for each frame of the image, mark the time points when the abnormal behavior or micro-expression starts and ends; behavior type annotation, specifically: assign a category to each abnormal behavior, such as "fatigue driving", "distracted operation", "angry expression", etc. Ensure that each category has a clear definition and examples to ensure the consistency and accuracy of the annotation; score annotation, specifically annotated by expert experience method; image data annotation includes region annotation, specifically: box out the area containing abnormal behavior or micro-expression on the image; key point annotation, specifically: annotate the key points of the face, such as the position of the eyes, mouth, and eyebrows, to help the model capture subtle changes in facial expressions; audio data annotation includes abnormal type annotation, specifically: abnormal sounds in the audio such as "rapid breathing" and "screaming", etc.; intensity scoring, specifically: evaluate the intensity of abnormal sounds based on the loudness, duration and frequency characteristics of the sound, and give corresponding scores. This process is also obtained through expert scoring; the above annotations are completed by Labelbox, VGGImage Annotator (VIA) or LabelImg annotation tools;

[0111] D3. Construct a behavior and micro-expression recognition model, which includes a video coding layer, an image coding layer, an audio coding layer, an attention fusion layer, an output layer 1, and an output layer 2. Input the acquired original abnormal behavior and micro-expression video data into the video coding layer in the behavior and micro-expression recognition model, output the abnormal behavior and micro-expression time change characteristics, input the annotated behavior and micro-expression image data into the image coding layer, output the abnormal behavior and micro-expression static characteristics, input the annotated audio data into the audio coding layer, and output the corresponding abnormal behavior audio characteristics; further, in this embodiment, the video coding layer is constructed by pre-trained SlowFast Networks and a bidirectional TCN network; the image coding layer is constructed by a pre-trained lightweight VIT model; and the audio coding layer is constructed by AudioSet Pre-TrainedModels or OpenL3;

[0112] D4. Input the acquired abnormal behavior and micro-expression time change features, abnormal behavior and micro-expression static features, and abnormal behavior audio features into the attention fusion layer constructed by the collaborative attention network to fuse and obtain abnormal behavior and micro-expression fusion features;

[0113] D5. Input the acquired abnormal behavior and micro-expression fusion features into output layer 1 and output layer 2 at the same time, and obtain the abnormal behavior type and the corresponding score respectively. and abnormal micro-expression types and corresponding scores ;

[0114] D6. According to the corresponding abnormal behaviors and micro-expression types and corresponding scores, set the cross entropy loss function and the corresponding loss threshold, train the D1-D5 process, and obtain the trained behavior and micro-expression recognition model;

[0115] Specifically, the specific steps of the behavior recognition unit monitoring the driver's micro-expressions and abnormal behavior during driving include:

[0116] D7. The trained behavior and micro-expression recognition model is built into the behavior recognition unit. During the driving of the transport vehicle, the driver's behavior, micro-expression and audio data are obtained in real time through the configured visual sensor;

[0117] D8. Input the acquired driving behavior, micro-expression and audio data into the behavior recognition unit to obtain the corresponding abnormal behavior and micro-expression type and the corresponding score and ;

[0118] D9. Set abnormal behavior score threshold and abnormal micro-expression score threshold ,when and When the driver drives normally, or It determines when the driver's driving is abnormal, and feeds back the acquired abnormal behavior and micro-expression types to the central control platform, generates corresponding processing strategies, and transmits them to the corresponding vehicle drivers in real time, and issues early warnings for abnormalities passing through the on-board sensors of transport vehicles.

[0119] The early warning and response unit is used to monitor and identify the abnormal monitoring results of the abnormal trajectory recognition unit and the behavior recognition unit, and to issue an early warning of the identified abnormal type through the vehicle-mounted equipment through the early warning and response measurement processing library built into the central control platform. At the same time, according to the corresponding abnormal type results, the corresponding emergency response processing strategy is generated and issued to the corresponding abnormal transport vehicle through the matching and generation algorithm; the emergency response processing strategy in this embodiment includes one-click alarm and automatic sending of the accident location to the rescue center, recording data before and after the accident, and assisting in post-analysis and responsibility determination;

[0120] The vehicle health management module is used to regularly monitor transport vehicles and implement preventive maintenance to extend the service life of transport vehicles; the vehicle health management module includes a regular inspection unit, a fault prediction unit, and a maintenance report generation unit;

[0121] A periodic inspection unit, used to perform periodic inspections on transport vehicles pre-registered in the transport vehicle information database to obtain attribute data of transport vehicle components;

[0122] A fault prediction unit is used to obtain the type and probability of a fault occurring during the transportation process of the transport vehicle in the next transportation cycle through a prediction algorithm based on the property data of the transport vehicle components obtained by regular detection and the running status score of the transport vehicle obtained by the abnormal trajectory recognition unit, and determine whether to allocate and execute tasks based on the probability of the fault occurring; further, the prediction algorithm in this embodiment includes a support vector machine, a random forest and a decision tree; further, in this embodiment, a fault prediction unit sets a fault probability threshold. When the probability of a fault occurring during the transportation process of the transport vehicle in the next transportation cycle obtained by the prediction algorithm exceeds the fault probability threshold, the corresponding probability and fault type are fed back to the central control platform, otherwise the predicted probability and fault type are not fed back to the central control platform. In this embodiment, the fault probability threshold is 0.7;

[0123] The maintenance report generation unit is used to regularly generate maintenance recommendation reports of corresponding fault types using text or video generation algorithms based on the acquired transport vehicle component attribute data, transport vehicle operation status scores, and predicted fault types and probabilities, to assist in the regular maintenance of transport vehicles and reduce the safety risks of transport vehicles. Furthermore, the text or video generation algorithms in this embodiment include: GPT (Generative Pre-trained Transformer) series, T5, and VideoGAN;

[0124] Example 2

[0125] See also Figure 3 Another embodiment provided by the present invention is: a corresponding unit workflow of a transportation vehicle safety management and control system specifically includes:

[0126] First, a transport vehicle information database is constructed through the vehicle information database unit deployed in the central control platform, and the attributes and qualification information of new pre-registered transport vehicles are reviewed in real time. At the same time, the pre-registration and update and save of transport vehicle information are completed for the transport vehicles that have passed the review; when the central control platform issues a task, the transport vehicle identity is confirmed by the label recognition unit. After the identity is confirmed, the prediction result obtained by the corresponding transport vehicle monitoring and maintenance unit is fed back to the central control platform to determine whether to issue task instruction information to the corresponding transport vehicle. When the corresponding transport vehicle receives the task instruction information after determination, it is determined whether the transport vehicle is allowed to enter and exit the specified area according to the transport vehicle label information and task instruction information through the entry and exit determination rules in the entry and exit control unit;

[0127] Second, when it is determined that the transport vehicle is allowed to enter and exit the specified area, the transport safety monitoring module is started synchronously, and the operation attribute data, position and trajectory data of the transport vehicle performing the task and the video and image data of the driver during driving are monitored in real time through the vehicle monitoring unit, and the acquired data is pre-processed and saved through the storage unit, and the saved operation attribute data and position and trajectory data of the transport vehicle are input into the abnormal trajectory recognition unit, and the abnormal driving trajectory of the transport vehicle and the running status score of the transport vehicle are monitored in real time. At the same time, the video and image data of the driver during driving are input into the behavior recognition unit, and the abnormal behavior status of the driver during driving is monitored; the abnormal driving trajectory of the transport vehicle, the running status score of the transport vehicle and the abnormal behavior status are input into the early warning and response unit for early warning response, and the corresponding emergency response processing strategy is generated and sent to the corresponding abnormal transport vehicle for abnormal processing;

[0128] Third, the transport vehicles are regularly inspected through the periodic inspection unit, and the inspection results are transmitted to the fault prediction unit to predict the type and probability of faults that will occur during the next transport cycle of the transport vehicle, and the prediction results are fed back to the central control platform. The central control platform determines whether to issue task instructions to the corresponding transport vehicles based on the prediction results. At the same time, the regular transport vehicle inspection results and the transport vehicle prediction results are transmitted to the maintenance report generation unit to generate a maintenance recommendation report for the corresponding fault type, and the generated maintenance recommendation report is displayed through the interactive interface of the central control platform to assist in the regular maintenance of the transport vehicles and reduce the safety risks of the transport vehicles.

[0129] The transport vehicle safety management and control system realizes comprehensive management and control of transport vehicles from pre-registration to task execution to regular maintenance through refined workflow design. The system not only improves the accuracy and timeliness of transport vehicle information management, but also effectively prevents potential safety hazards through real-time monitoring and intelligent early warning, ensuring the safety and smoothness of the transportation process. In addition, the system's regular detection and fault prediction functions provide a scientific basis for transport vehicle maintenance, making maintenance work more accurate and efficient, and further reducing the safety risks of transport vehicles.

[0130] Example 3

[0131] See also Figure 4 Another embodiment provided by the present invention is a method for safety management and control of transport vehicles, the specific steps of which include:

[0132] S1. Obtain the attributes and qualification text information of transport vehicles, build a transport vehicle information database through natural language algorithms, and configure a qualification review model to review the transport vehicle information added in real time;

[0133] S2. According to the constructed transport vehicle information database, the identity of the transport vehicle is confirmed through the electronic tag and matching recognition algorithm. According to the confirmation result and the task instruction issued by the central control platform, the entry and exit judgment rules are used to judge whether the transport vehicle can enter and exit the specified area, and the task is executed;

[0134] S3. When it is determined that the transport vehicle is capable of performing the task, the operating attribute data, position and trajectory data of the transport vehicle performing the task and the video and image data of the driver during driving are simultaneously acquired through real-time monitoring by sensors;

[0135] S4. Based on the operation attribute data, location and trajectory data of the transport vehicle, the abnormal driving trajectory of the transport vehicle and the operation status score of the transport vehicle are obtained through the trajectory recognition algorithm and the comprehensive evaluation algorithm. At the same time, based on the video and image data of the driver during driving, the micro-expression and abnormal behavior status of the driver during driving are obtained through the behavior and micro-expression recognition model;

[0136] S5. Based on the acquired abnormal driving trajectory and operating status scores of the transport vehicle and the micro-expressions and abnormal behavior of the driver during driving, the warning and response measurement processing library built into the central control platform is used to issue warnings for the identified abnormal types through the on-board equipment. At the same time, based on the corresponding abnormal type results, the corresponding emergency response processing strategy is generated and issued to the corresponding abnormal transport vehicle through matching and generation algorithms;

[0137] S6. Perform regular inspections on pre-registered transport vehicles to obtain vehicle component attribute data, and use a prediction algorithm to predict the type and probability of failure of the transport vehicle in the next cycle based on the vehicle component attribute data and the transport vehicle operation status score, and feed back the acquired failure type and probability to the central control platform to determine whether to issue a task instruction;

[0138] S7. Based on the acquired transport vehicle component attribute data, transport vehicle operation status score and predicted fault type and probability, a text or video generation algorithm is used to regularly generate maintenance recommendation reports corresponding to the fault type to assist in the regular maintenance of transport vehicles and reduce the safety risks of transport vehicles.

[0139] Example 4

[0140] A computer-readable storage medium stores computer instructions, which, when executed, execute a method for safety management and control of transport vehicles.

[0141] The embodiments of the present invention are described above in conjunction with the accompanying drawings, but the present invention is not limited to the above-mentioned specific implementation modes, which are merely illustrative rather than restrictive. Under the guidance of the present invention, ordinary technicians in the field may also change, modify, replace and modify the above-mentioned embodiments without departing from the scope of protection of the purpose of the present invention and the claims, and all of these are within the protection of the present invention.

[0142] If the disclosed technical solution involves personal information, the product using the disclosed technical solution has clearly informed the personal information processing rules and obtained the individual's voluntary consent before processing the personal information. If the disclosed technical solution involves sensitive personal information, the product using the disclosed technical solution has obtained the individual's separate consent before processing the sensitive personal information, and at the same time meets the "explicit consent" requirement. For example, on personal information collection devices such as cameras, clear and prominent signs are set to inform that the personal information collection scope has been entered and personal information will be collected. If the individual voluntarily enters the collection scope, it is deemed that he or she agrees to the collection of his or her personal information; or on the device that processes personal information, when the personal information processing rules are notified by obvious signs / information, the individual's authorization is obtained through pop-up information or by asking the individual to upload his or her personal information; among them, the personal information processing rules may include information such as the personal information processor, the purpose of personal information processing, the processing method, and the type of personal information processed.

Claims

1. A transport vehicle safety management and control system, characterized in that: include: Central control platform, vehicle access module and transportation safety monitoring module; The central control platform is used for the deployment and operation of all modules and units in the security management and control system, as well as the uploading of operation results and the allocation and issuance of execution task instructions; The vehicle access module includes a vehicle information database unit, a tag identification unit and an access control unit; The vehicle information database unit is used to construct a transport vehicle information database using a natural language algorithm based on the acquired transport vehicle attributes and qualification text information, and configure a qualification review model at the transport vehicle information database interface position to conduct qualification review on the transport vehicle information added in real time, and complete pre-registration and transport vehicle information storage for the transport vehicles that have passed the review; the label recognition unit is used to match and identify the transport vehicle label information and the pre-registered information in the transport vehicle information database based on the electronic label installed on the transport vehicle through a matching recognition algorithm to confirm the identity of the transport vehicle; the entry and exit control unit is used to construct entry and exit discrimination rules, and automatically determine whether the transport vehicle is allowed to enter and exit the specified area based on the matching recognition results of the label recognition unit and the task instructions issued by the central control platform through the entry and exit discrimination rules; The transport safety monitoring module includes a transport vehicle monitoring unit, an abnormal trajectory recognition unit and a behavior recognition unit; The vehicle monitoring unit is used to monitor and obtain the operation attribute data, position and trajectory data of the transport vehicle that is performing the task in the transport vehicle information database in real time through the deployed integrated sensors, and the video and image data of the driver during the driving process; The abnormal trajectory identification unit is used to monitor the abnormality of the driving trajectory of the transport vehicle and the running status score of the transport vehicle in real time through a trajectory identification algorithm and a comprehensive evaluation algorithm according to the real-time position and trajectory data and running attribute data of the transport vehicle; The behavior recognition unit is used to monitor the driver's micro-expressions and abnormal behavior status during driving in real time through the behavior and micro-expression recognition model based on the video and image data of the driver during driving. The specified area includes a parking area and a working area, and the specific implementation steps of the access control unit include: B1. When transport vehicle A is not performing a task, the label information of transport vehicle A is identified by the label identification unit to obtain the corresponding label information of transport vehicle A, and a matching algorithm is used to match the obtained label information of transport vehicle A with the pre-registered information in the transport vehicle information database. , the specific formula is: ， in, , and It represents the similarity between the license plate number, vehicle model and driver information corresponding to the label information obtained by identifying the transport vehicle A and the pre-registered information in the transport vehicle information database, It indicates identifying the transport vehicle A and obtaining the corresponding license plate number information in the label information. Indicates the pre-registered license plate number information in the transport vehicle information database. It indicates identifying the transport vehicle A and obtaining the corresponding vehicle model information in the label information. Indicates the vehicle model information pre-registered in the transport vehicle information database. It indicates identifying the transport vehicle A and obtaining the corresponding driver information in the label information. Represents the driver information pre-registered in the transport vehicle information database; represents the corresponding weighting coefficient; B2. Set the tag information matching threshold , the obtained matching degree and For comparison, When , the transport vehicle A is allowed to enter and exit the parking area, but is prohibited from entering and exiting the working area. When the vehicle is parked, only transport vehicle A is allowed to leave the parking area, and is prohibited from entering the parking area and entering and exiting the work area; B3: When transport vehicle A performs a task, the label information of transport vehicle A is judged through B1 and B2. If it satisfies If the condition is not met, the task instruction of transport vehicle A is judged again. Then they are directly prohibited from entering the work area; The specific steps of performing secondary determination on the task instruction of the transport vehicle A include: B4, obtain the task instruction information of transport vehicle A through sensors, and use visual sensors and weight sensors to obtain the type and weight information of the goods transported by transport vehicle A at the current moment; B5. Based on the task instruction information obtained in B4 and the type and weight information of the goods transported by vehicle A at the current moment, a task instruction verification process formula is constructed through Boolean functions. The obtained task instruction information of transport vehicle A is compared and judged with the information in the preset task list of the central control platform. When all the comparisons and judgments are passed, transport vehicle A is allowed to enter and exit the working area and parking area. The specific formula of the task instruction verification process is: ， in, It means to verify whether the task instruction of transport vehicle A exists in the preset task list. It verifies whether the task execution time of transport vehicle A is within the allowed time window. It verifies whether the mission purpose in the mission instruction of transport vehicle A matches the current destination. It verifies whether the authority level of transport vehicle A matches the task requirements, that is, whether the current transport vehicle A meets the corresponding rated cargo capacity in the current task requirements. It verifies whether the cargo type currently transported by transport vehicle A matches the cargo type in the task instruction. It verifies whether the current cargo weight of transport vehicle A is consistent with the cargo weight in the task instruction. Represents the "and" operation in the logical operator; B6. If one of the comparisons fails, vehicle A is prohibited from entering or exiting the work area, but is allowed to enter or exit the parking area; The specific steps of constructing the entry and exit discrimination rules in the entry and exit control unit include: According to the B1-B6 process, the entry and exit discrimination rules are constructed, and the constructed entry and exit discrimination rules are built into the entry and exit control unit to discriminate the transport vehicles entering and leaving the specified area. The specific formula of the entry and exit discrimination rules is: ， Where A=S&A W means that transport vehicle A is allowed to enter and exit the parking area, but is prohibited from entering and exiting the working area; A S&A W means that transport vehicle A is prohibited from entering or leaving the parking area and the working area. Represents the process formula verified by task instructions Carry out comparison and judgment, and all the judgments pass. Represents the process formula verified by task instructions Perform a comparison and judgment to determine if at least one of the verifications failed, A=S&A W means that the transport vehicle A is allowed to enter and exit the parking area and the working area; A means the transport vehicle A, S means the parking area, W means the working area, T means that v is judged to be true, that is, True, and F means that v is judged to be false, that is, v is False; The task instruction issued by the central control platform includes the starting point, transportation time, destination and optimal driving path trajectory of the transport vehicle. The optimal driving path trajectory is automatically generated by the central control platform through a path planning generation algorithm according to the task instruction information issued and the road condition information obtained in real time, and transmitted to the corresponding transport vehicle; the specific steps of the abnormal trajectory recognition unit monitoring the abnormal driving trajectory of the transport vehicle through the trajectory recognition algorithm include: C1. Set the trajectory deviation threshold , and monitor the process trajectory data of the transport vehicle traveling along the optimal driving path in real time, obtain the coordinate information of each position in the process trajectory data and the position information in the optimal driving trajectory in the task instruction, and calculate the position difference through the spatial distance formula ,in It represents the distance between the i-th position coordinate in the real-time monitoring process trajectory data and the corresponding position coordinate in the optimal driving trajectory; C2, the position difference obtained Threshold for deviation from track For comparison, when N consecutive position differences in the real-time monitoring process trajectory data are greater than the trajectory deviation threshold When , it is determined that the vehicle position is offset; C3. When the position is offset, the current position coordinates after the offset, the current road condition information and the target point position are monitored, and a path planning algorithm is used to determine whether the current position coordinates after the offset can reach the target point position within the transportation time length; C4. If it is determined that the transport vehicle cannot reach the target location or cannot reach the target location within the transport time, the track deviation is determined to be abnormal, and a track deviation abnormality warning is sent to the vehicle; C5. When it is determined that the transport vehicle can reach the target point within the transport time, a new optimal driving path trajectory is generated according to the current position coordinates after the offset and the target point position, and sent to the corresponding transport vehicle. The corresponding transport vehicle is continuously monitored on the new optimal driving path trajectory to determine whether a secondary trajectory offset abnormality warning occurs.

2. A transport vehicle safety control system as claimed in claim 1, characterized in that: The specific workflow of the vehicle information database unit includes: A1. Build an integrated model of image text extraction and entity-relationship joint extraction based on the CTPN pre-trained sub-model, the Chinese pre-trained BERT sub-model, the BILSTM sub-model and the CRF sub-model, obtain vehicle information text data and driver information data, and use the obtained vehicle information text data to fine-tune the integrated model of image text extraction and entity-relationship joint extraction to obtain the trained integrated model of image text extraction and entity-relationship joint extraction; A2. The trained integrated model of image text extraction and entity-relationship joint extraction is built into the Neo4j graph database. When the driver uploads the corresponding text information, the built-in integrated model of image text extraction and entity-relationship joint extraction is used to perform image or text recognition and entity relationship extraction on the information uploaded by the driver to obtain vehicle information and driver entity relationship data; A3. Build and train the qualification review model based on the Chinese pre-trained BERT sub-model and the cross-attention sub-model, embed the obtained qualification review model into the Neo4j graph database interface, match the vehicle information and driver entity relationship data obtained in A2 with the qualification review conditions configured in the Neo4j graph database, and obtain the matching similarity ; A4. Set the audit match threshold. When the matching similarity is greater than or equal to the audit match threshold, the audit is automatically passed, and the acquired vehicle information and driver entity relationship data are automatically saved in the transport vehicle information database built by the Neo4j graph database to complete the vehicle information storage and driver information pre-registration.

3. A transport vehicle safety control system as claimed in claim 2, characterized in that: The specific steps of the behavior recognition unit monitoring the driver's micro-expressions and abnormal behavior during driving include: D1. Obtain the historical driver's abnormal behavior and micro-expression video and image data during driving, and perform frame processing on the acquired abnormal video data through the OpenCV library to obtain abnormal micro-expression and behavior image data, and obtain the corresponding audio data when the abnormal behavior and micro-expression occur from the abnormal behavior and micro-expression video data through audio extraction technology; D2. Annotate the above-obtained abnormal behavior and micro-expression video, image and audio data with corresponding abnormal labels, and annotate the obtained abnormal behavior and micro-expression video and image data with abnormal behavior types and scores. At the same time, annotate the corresponding abnormal types for the audio data, and annotate the audio data with scores based on the intensity of the abnormal sounds in the audio data; D3. Construct a behavior and micro-expression recognition model, input the acquired original abnormal behavior and micro-expression video data into the video coding layer of the behavior and micro-expression recognition model, output the abnormal behavior and micro-expression time change features, input the annotated behavior and micro-expression abnormal image data into the image coding layer, output the abnormal behavior and micro-expression static features, input the annotated audio data into the audio coding layer, and output the corresponding abnormal behavior audio features; D4. Input the acquired abnormal behavior and micro-expression time change features, abnormal behavior and micro-expression static features, and abnormal behavior audio features into the attention fusion layer constructed by the collaborative attention network for fusion, and obtain abnormal behavior and micro-expression fusion features; D5. Input the acquired abnormal behavior and micro-expression fusion features into output layer 1 and output layer 2 at the same time, and obtain the abnormal behavior type and the corresponding score respectively. and abnormal micro-expression types and corresponding scores ; D6. According to the corresponding abnormal behaviors and micro-expression types and corresponding scores, set the cross entropy loss function and the corresponding loss threshold, repeat the D1-D5 process, train the behavior and micro-expression recognition model, and obtain the trained behavior and micro-expression recognition model.

4. A transport vehicle safety control system as claimed in claim 3, characterized in that: The specific steps of the behavior recognition unit monitoring the driver's micro-expressions and abnormal behavior during driving also include: D7. The trained behavior and micro-expression recognition model is built into the behavior recognition unit. During the driving of the transport vehicle, the driver's behavior, micro-expression and audio data are obtained in real time through the configured visual sensor; D8. Input the acquired driving behavior, micro-expression and audio data into the behavior recognition unit to obtain the corresponding abnormal behavior and micro-expression type and the corresponding score and ; D9. Set abnormal behavior score threshold and abnormal micro-expression score threshold ,when and When the driver drives normally, or When the driver's driving is abnormal, the abnormal behavior and micro-expression types obtained are fed back to the central control platform, a corresponding processing strategy is generated, and it is transmitted to the corresponding vehicle driver in real time, and an early warning is issued for the abnormality passing through the on-board sensors of the transport vehicle.

5. A method for transport vehicle safety management and control, which is implemented based on a transport vehicle safety management and control system according to any one of claims 1 to 4, characterized in that: The specific steps include: S1. Obtain the attributes and qualification text information of transport vehicles, build a transport vehicle information database through natural language algorithms, and configure a qualification review model to review the transport vehicle information added in real time; S2. According to the transport vehicle information database, the transport vehicle is identified through the electronic tag and matching recognition algorithm. According to the confirmation result and the task instruction issued by the central control platform, the entry and exit discrimination rules are used to determine whether the transport vehicle can enter and exit the specified area, and the task is executed; S3. When it is determined that the transport vehicle is capable of performing the task, the operating attribute data, position and trajectory data of the transport vehicle performing the task and the video and image data of the driver during driving are simultaneously acquired through real-time monitoring by sensors; S4. Based on the operation attribute data, location and trajectory data of the transport vehicle, the abnormal driving trajectory of the transport vehicle and the operation status score of the transport vehicle are obtained through the trajectory recognition algorithm and the comprehensive evaluation algorithm. At the same time, based on the video and image data of the driver during driving, the micro-expression and abnormal behavior status of the driver during driving are obtained through the behavior and micro-expression recognition model; S5. Based on the acquired abnormal driving trajectory and operating status scores of the transport vehicle and the driver's micro-expressions and abnormal behavior during driving, the warning and response measurement and processing library built into the central control platform is used to issue warnings for the identified abnormal types through the on-board equipment. At the same time, based on the corresponding abnormal type results, the corresponding emergency response processing strategy is generated and issued to the corresponding abnormal transport vehicle through matching and generation algorithms.

6. A computer-readable storage medium, characterized in that: Computer instructions are stored thereon, and when the computer instructions are executed, a method for safety management and control of transport vehicles as described in claim 5 is executed.

Images