Remote monitoring method and system for large intelligent driving beam car

Abstract

This invention relates to a remote monitoring method and system for a large intelligent driving beam transport vehicle, belonging to the field of vehicle remote monitoring and intelligent control technology. It includes acquiring the beam transport vehicle's driving information and forming a full-coverage real-time information stream; marking the full-coverage real-time information stream into three types of events—high-weight events, medium-weight events, and low-weight events—based on weight information, prioritizing and processing driving information with high weight levels; issuing corresponding operation instructions based on real-time events; and changing the weight level and reducing risk warnings for events that have already undergone specific operations, in conjunction with the real-time information stream. This invention improves the efficiency and completeness of remote monitoring by acquiring driving information and forming a real-time information stream, monitoring the real-time information stream, weight marking, and closed-loop monitoring processing, significantly enhancing the intelligence level of the remote monitoring system in information processing, risk identification, and emergency response.

Description

Technical Field

[0001] This invention relates to the field of vehicle remote monitoring and intelligent control technology, and in particular to a remote monitoring method and system for a large intelligent driving beam transport vehicle. Background Technology

[0002] With the continuous advancement of infrastructure construction, the demand for transporting precast bridge components has increased significantly, and large beam transport vehicles play a core transportation role in modern bridge engineering. These vehicles are large and heavy, and their operating routes often involve non-standard road conditions such as construction roads and temporary bridges, placing higher demands on driving safety and handling performance. To improve the transportation efficiency and safety level of beam transport vehicles, intelligent driving and remote monitoring technologies are gradually being applied to the scheduling and management systems of beam transport vehicles. Currently, intelligent beam transport vehicle systems typically include an autonomous driving module, an environmental perception module, a communication module, and a remote monitoring center. Relying on sensor fusion and real-time communication, remote control and status management of the beam transport vehicle are achieved. However, in practical applications, existing intelligent systems still face several key challenges: 1. Insufficient information collection dimensions make it difficult to achieve comprehensive perception of the beam transport vehicle's status and environment, resulting in incomplete data held by the monitoring center; 2. Real-time data lacks effective hierarchical classification and processing, and important information is not highlighted; 3. The lack of a closed-loop feedback mechanism for high-risk events may lead to untimely responses to abnormal situations; 4. The system's high dependence on manual operation easily leads to problems such as misoperation and monitoring oversights.

[0003] Therefore, there is an urgent need to build a remote monitoring method and system with full-coverage information perception, risk level identification, two-way interactive control and response closed-loop mechanism to improve the operational safety and efficiency of intelligent beam transport vehicles in complex construction environments. Summary of the Invention

[0004] In order to solve the above-mentioned technical problems, the present invention provides a remote monitoring method and system for a large intelligent driving beam transport vehicle.

[0005] This invention adopts the following technical solution: a remote monitoring method for a large intelligent driving beam transport vehicle, comprising:

[0006] Acquire the beam transport vehicle's driving information and form a comprehensive real-time information stream;

[0007] The real-time information stream with full coverage is marked into three types of events based on weight information: high-weight events, medium-weight events, and low-weight events. Events with high weight are given priority in notification and processing.

[0008] Issue corresponding operation instructions based on real-time events;

[0009] For events that have already been processed, combine real-time information flow to change weight levels and reduce risk warnings;

[0010] Among them, the real-time information stream with full coverage will be marked into three types of events based on weight information: high-weight events, medium-weight events, and low-weight events.

[0011] Set up a rules engine to fully cover the real-time information flow and identify event types and impact levels according to the rules engine;

[0012] Different weight levels are assigned based on the type of event and the degree of its impact.

[0013] In some embodiments, the driving information includes real-time data of vehicle condition parameters, driving status, and surrounding images.

[0014] In some embodiments, the vehicle condition parameters include: the operating status of the power system, braking unit, power module, and suspension device of the beam transport vehicle;

[0015] Driving status includes: speed, acceleration, angular velocity, heading, and positioning data of the beam transport vehicle;

[0016] Real-time data of the surrounding images includes images covering the forward, lateral, and blind spot areas of the beam transport vehicle.

[0017] In some embodiments, a rule engine is configured to fully cover the real-time information stream by identifying event types and impact levels according to the rule engine, including:

[0018] Define multi-dimensional rules to form the rule base conditions, which include event type, threshold conditions, duration of time, and degree of impact.

[0019] Receive a full-coverage real-time information stream;

[0020] The received full-coverage real-time information stream is compared with the rules base conditions to identify the event type and determine the degree of impact.

[0021] In some embodiments, different weight levels are assigned based on the event type and the degree of impact, according to the following formula:

[0022]

[0023] in:

[0024] Weighting coefficient;

[0025] Final weighted score;

[0026] Event type;

[0027] : Duration score, the longer the duration, the higher the score;

[0028] Degree of impact;

[0029] : Feedback processing status, +10 for unprocessed, -5 for processed;

[0030] Defined as a high-weight event, Defined as a medium-weighted event, Defined as a weighted event.

[0031] In some embodiments, assigning different weight levels based on event type and impact is accomplished by a trained feedforward neural network, which includes:

[0032] The input layer is used to input the event type, threshold conditions, duration, and impact level.

[0033] The hidden layer processes the input signal and extracts features, and performs non-linear transformations through an activation function; the activation function used is the Sigmoid function, which compresses the result to between 0 and 1.

[0034] The output layer is used to output the weight scores corresponding to the events.

[0035] In some embodiments, issuing corresponding operation instructions based on real-time events includes:

[0036] When low-weight events are identified:

[0037] The beam transport vehicle responds autonomously, only recording the event without triggering an active alarm; the beam transport vehicle then adjusts autonomously; if a low-weight event continues to be displayed for more than the preset duration, it is upgraded to a medium-weight event.

[0038] When a medium-weighted event is identified:

[0039] For events that are pre-set to be handled autonomously by the intelligent driving system of the beam transport vehicle, the system automatically triggers action responses to reduce the speed and adjust the steering of the beam transport vehicle;

[0040] For events that are pre-defined as having an undetermined root cause and potential risks and require manual review or intervention, the monitoring personnel are prompted to select "ignore," "downgrade," or "upgrade to a high-weight event."

[0041] If a medium-weighted event continues to be displayed for a duration exceeding the preset duration, it will be automatically promoted to a high-weighted event. If it returns to normal, it will be marked as "resolved" and its weight level will be reduced.

[0042] When high-weight events are identified:

[0043] For events that are pre-set to be handled autonomously by the intelligent driving system of the beam transport vehicle, the system automatically triggers action responses to reduce the speed and adjust the steering of the beam transport vehicle;

[0044] For events that are pre-defined as having an undetermined root cause and potential risks, requiring manual review or intervention, a high-priority alert is issued to prompt monitoring personnel to make a timely assessment and issue control commands.

[0045] In some embodiments, combining events that have already undergone specific actions with a comprehensive real-time information stream, changing weight levels and reducing risk alerts includes:

[0046] Once a high-weight event is identified and a response operation is triggered, a closed-loop task node is automatically generated to record the event identifier, initial weight, instruction content, execution feedback, and subsequent tracking status.

[0047] During the operation of the beam transport vehicle, the feedback channel of the beam transport vehicle is continuously monitored, and the weight level of the event is dynamically adjusted according to the changes in the driving status of the beam transport vehicle.

[0048] If a closed-loop task fails to respond within the specified time, its processing priority will be increased, and a reminder will be sent to the monitoring personnel.

[0049] In some embodiments, combining events that have already undergone specific operations with a full-coverage real-time information stream, changing weight levels and reducing risk warnings also includes: assigning a unique number to each high-weight event; after each high-weight event is identified and a response operation is triggered, the relevant control instructions, feedback information, status adjustments and final processing results are all recorded for subsequent retrospective review and analysis.

[0050] In some embodiments, the weighting of an event is adjusted using the following formula:

[0051]

[0052] in:

[0053] : The initial weight score of the information item at time t;

[0054] Operational feedback;

[0055] Real-time status updates;

[0056] : Operational feedback dominance coefficient, indicating the degree of influence of feedback on weight adjustment;

[0057] Real-time status dominant coefficient, indicating the degree of influence of new data on weight adjustment;

[0058] + =1, ensuring the weight update results are normalized. > This indicates that the dominant weight of the processing result has decreased; if the feedback is good, The weighting is reduced; if potential risks still exist, Maintain weight level prompts.

[0059] In some embodiments, after changing the weight level of an event, a new determination is made as to whether to continue prompting the corresponding operation instruction or to reduce the weight level. The determination process is represented by the following formula:

[0060]

[0061] in:

[0062] Lower the trigger threshold;

[0063] Warning trigger threshold;

[0064] : Positive effect factor of feedback after response;

[0065] : The residual risk factor of the event in the next time step state.

[0066] A remote monitoring system for a large intelligent driving beam transport vehicle, comprising a method for remotely monitoring the large intelligent driving beam transport vehicle, including:

[0067] The data acquisition module is used to continuously collect driving information during the operation of the beam transport vehicle, including real-time data on vehicle condition parameters, driving status, and surrounding images.

[0068] The full-coverage real-time information stream generation module is used to integrate the driving information collected by the data acquisition module into the data channel to form a structured and consistent full-coverage real-time information stream.

[0069] The full-coverage real-time information stream weight training and allocation module is used to automatically label information with different weight levels in the full-coverage real-time information stream of the beam transport vehicle, and to give priority to and process the full-coverage real-time information stream of the beam transport vehicle with higher weight.

[0070] The remote monitoring module is used to receive the full-coverage real-time information stream from the beam transport vehicle and issue corresponding operation instructions based on the real-time weight level information.

[0071] The monitoring closed-loop processing module is used to change the weight level and risk warnings based on the weight information that has already been processed, combined with the full-coverage real-time information flow.

[0072] In some embodiments, the remote monitoring module includes an information display interface, a command issuance system, and a two-way voice communication function, used to establish a real-time perception and control closed loop with the beam transport vehicle.

[0073] In some embodiments, the information display interface adopts a graphical interface, which classifies and presents the real-time information stream of the vehicle in real time according to a preset weight level; high-weight events are highlighted by color coding and visual hierarchy, accompanied by audible and visual alarm prompts, so as to facilitate timely identification by monitoring personnel; each abnormal information includes trigger time, event type, vehicle number, geographical location, status parameters and image snapshot, and monitoring personnel can click to view details and select response operations.

[0074] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0075] By acquiring the driving information of the intelligent driving beam transport vehicle and forming a full-coverage real-time information stream, the system simultaneously collects the vehicle's operating status, vehicle condition parameters, and on-site images. This unified, full-coverage real-time information stream enables visualized monitoring and status identification throughout the entire process. By labeling the full-coverage real-time information stream with different weight levels, and prioritizing and processing driving information with higher weight levels, the system constructs event recognition rules and a weighted classification model. This allows for the priority identification and prominent alerting of high-risk information, significantly enhancing the timeliness of anomaly responses. Through the remote monitoring module, corresponding operation commands are issued based on the real-time weight level information, enabling the monitoring platform to generate control commands. The system enables rapid issuance of commands, improving the immediacy and efficiency of vehicle information transmission between operators and the beam transport vehicle. By combining weighted information of already performed operations with real-time information flow, it changes weight levels and reduces risk warnings, achieving closed-loop tracking and adaptive risk management of abnormal response events. In summary, this invention improves the efficiency of remote monitoring and the integrity of information processing through full-coverage real-time information flow monitoring, weight marking, and closed-loop monitoring processing. It significantly enhances the intelligence level of the remote monitoring system in information processing, risk identification, and emergency response, thereby ensuring the safety and efficiency of large intelligent driving beam transport vehicles in complex construction environments. Attached Figure Description

[0076] Figure 1 The flowchart shows the remote monitoring method and system for the large intelligent driving beam transport vehicle in this invention.

[0077] Figure 2 This is a flowchart illustrating the specific implementation of the remote monitoring method and system for the large intelligent driving beam transport vehicle of the present invention.

[0078] Figure 3 This is a schematic diagram of the data acquisition process for the remote monitoring method and system of the large intelligent driving beam transport vehicle in this invention.

[0079] Figure 4 This is a flowchart of the closed-loop control process for high-weighted events in the remote monitoring of the large intelligent driving beam transport vehicle in this invention;

[0080] Figure 5 This is a schematic diagram of the information closed-loop processing of the remote monitoring method and system for the large intelligent driving beam transport vehicle of the present invention. Detailed Implementation

[0081] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0082] The remote monitoring system for a large intelligent driving beam transport vehicle provided by this invention is used to implement a remote monitoring method for the large intelligent driving beam transport vehicle, including:

[0083] The data acquisition module is used to continuously collect driving information during the operation of the beam transport vehicle, including real-time data on vehicle condition parameters, driving status, and surrounding images.

[0084] The full-coverage real-time information stream generation module is used to integrate the real-time data collected by the data acquisition module into the data channel to form a structured and consistent continuous full-coverage real-time information stream.

[0085] The full-coverage real-time information stream weight training and allocation module is used to automatically label information of different weight levels in the full-coverage real-time information stream of the beam transport vehicle according to the weight allocation algorithm and training data, and to give priority to and process the full-coverage real-time information stream of the beam transport vehicle with higher weight.

[0086] The remote monitoring module is used to receive the full-coverage real-time information stream from the beam transport vehicle and issue corresponding operation instructions based on the real-time weight level information.

[0087] The monitoring closed-loop processing module is used to change the weight level and risk warnings based on the weight information that has already been processed, combined with the full-coverage real-time information flow.

[0088] By setting up a data acquisition module, a full-coverage real-time information flow generation module, a full-coverage real-time information flow weight training and allocation module, a remote monitoring module, and a monitoring closed-loop processing module, the system collects driving information during the operation of the beam transport vehicle and forms a full-coverage information flow. Then, it automatically assigns different weight levels to the full-coverage information flow, prioritizing the processing of high-weight events. This achieves continuous monitoring of the beam transport vehicle during its operation, unifies and aggregates the monitoring data, and promptly identifies and processes high-weight events, improving the driving safety of the beam transport vehicle. Furthermore, it can continue monitoring after identifying high-weight events, adjusting the weight levels of events that are responded to and processed in a timely manner and those that are not, thus achieving closed-loop tracking and adaptive weight management of events.

[0089] The remote monitoring system for a large intelligent driving beam transport vehicle provided by this invention includes an information display interface, a command issuance system, and a two-way voice communication function, establishing a real-time perception and control closed loop between the system and the beam transport vehicle.

[0090] By setting up a remote monitoring module, not only is continuous monitoring of events achieved and communication between operators and beam transport vehicles improved, but it can also be used for manual confirmation and emergency dispatch in sudden situations or scenarios with ambiguous identification, thereby enhancing remote management capabilities.

[0091] The remote monitoring system for large intelligent driving beam transport vehicles provided by this invention features a graphical interface that categorizes and presents real-time vehicle condition, location, and image data according to preset weight levels. High-weight events are highlighted through color coding and visual hierarchy, accompanied by audible and visual alarms, facilitating timely identification by monitoring personnel. Each abnormal information includes trigger time, event type, vehicle number, geographical location, status parameters, and image snapshot. Monitoring personnel can click to view details and select response actions.

[0092] like Figures 1 to 5 As shown, the present invention provides a remote monitoring method for a large intelligent driving beam transport vehicle, comprising:

[0093] Acquire the beam transport vehicle's driving information and form a comprehensive real-time information stream;

[0094] The real-time information stream with full coverage is marked into three types of events based on weight information: high-weight events, medium-weight events, and low-weight events. Events with high weight are given priority in notification and processing.

[0095] Issue corresponding operation instructions based on real-time events;

[0096] For events that have already been processed, combine real-time information flow to change weight levels and reduce risk warnings;

[0097] Among them, the real-time information stream with full coverage will be marked into three types of events based on weight information: high-weight events, medium-weight events, and low-weight events.

[0098] Set up a rules engine to fully cover the real-time information flow and identify event types and impact levels according to the rules engine;

[0099] Different weight levels are assigned based on the type of event and the degree of its impact.

[0100] In a specific implementation, the data acquisition module relies on various types of sensors installed on the intelligent beam transport vehicle to continuously collect vehicle condition parameters, driving status, and surrounding image data during the vehicle's operation. This constructs a real-time information flow covering the entire process, all working conditions, and multiple perspectives, providing basic data support for weight recognition, remote response, and closed-loop control.

[0101] The lidar output includes three-dimensional coordinates (X, Y, Z) transformed to the geodetic coordinate system, as well as reflection intensity information; the high-precision inertial navigation system provides positioning coordinates and attitude angles. The lidar coordinate transformation process can be described by the following formula:

[0102]

[0103] In the process of marking the real-time information flow with full coverage into different weight levels and prioritizing and processing driving information with higher weight levels, the information of different types or sources should be automatically weighted according to the real-time information flow with full coverage of driving, and high-weight information should be prioritized and processed. The weighting function should comprehensively consider multiple characteristics such as information source, urgency, and reliability, and have the ability to adjust online adaptively.

[0104] By acquiring the driving information of the intelligent driving beam transport vehicle and forming a full-coverage real-time information stream, the system simultaneously collects the vehicle's operating status, vehicle condition parameters, and on-site images, fusing them to construct a unified real-time information stream. This enables full-process visual monitoring and status recognition. By marking the full-coverage real-time information stream with different weight levels, and prioritizing and processing driving information with higher weight levels, the system constructs event recognition rules and a weight classification model. This allows for the priority identification and prominent alerting of high-risk information, significantly enhancing the timeliness of anomaly response. Through the remote monitoring module, corresponding operation commands are issued based on the real-time weight level information, enabling the monitoring platform to generate control commands. The system enables rapid issuance of commands, improving the immediacy and efficiency of vehicle information transmission between operators and the beam transport vehicle. By combining weighted information of already performed operations with real-time information flow, it changes weight levels and reduces risk warnings, achieving closed-loop tracking and adaptive risk management of abnormal response events. In summary, this invention improves the efficiency of remote monitoring and the integrity of information processing through real-time information flow monitoring, weight marking, and closed-loop monitoring processing. It significantly enhances the intelligence level of the remote monitoring system in information processing, risk identification, and emergency response, thereby ensuring the safety and efficiency of large intelligent driving beam transport vehicles in complex construction environments.

[0105] The remote monitoring method for a large intelligent driving beam transport vehicle provided by this invention includes real-time data on vehicle condition parameters, driving status, and surrounding images.

[0106] The remote monitoring method for a large intelligent driving beam transport vehicle provided by the present invention includes vehicle condition parameters such as the operating status of the power system, braking unit, power module and suspension device of the beam transport vehicle.

[0107] Driving status includes: speed, acceleration, angular velocity, heading, and positioning data of the beam transport vehicle;

[0108] Real-time data of the surrounding images includes images covering the forward, lateral, and blind spot areas of the beam transport vehicle.

[0109] In a specific implementation, vehicle condition parameters are synchronously collected from multiple source nodes via the vehicle bus, and redundancy verification is performed in the main control unit; driving status is fused and acquired by combining the inertial measurement unit, satellite positioning system, and electronic compass; real-time data of surrounding images are collected by multi-view cameras, covering the vehicle's forward, lateral, and blind spot areas, and preliminary processing and feature extraction are performed at the edge end, and then uniformly merged into the data channel. After the middleware module completes the format standardization, time synchronization, and encoding compression, a structured and consistent continuous information flow is formed.

[0110] The remote monitoring method for large intelligent driving beam transport vehicles provided by this invention includes setting up a rule engine to fully cover real-time information flow and identify event types and impact levels according to the rule engine, including:

[0111] Define multi-dimensional rules to form the rule base conditions, which include event type, threshold conditions, duration of time, and degree of impact.

[0112] Receive a full-coverage real-time information stream;

[0113] The received full-coverage real-time information stream is compared with the rules base conditions to identify the event type and determine the degree of impact.

[0114] Specifically, based on the comparison results, and considering the event risk type, duration, impact, and handling status, three weight levels—high, medium, and low—are assigned.

[0115] The remote monitoring method for large intelligent driving beam transport vehicles provided by this invention assigns different weight levels based on event type and impact level, according to the following formula:

[0116]

[0117] in:

[0118] Weighting coefficient;

[0119] Final weighted score;

[0120] Event type;

[0121] : Duration score, the longer the duration, the higher the score;

[0122] Degree of impact;

[0123] : Feedback processing status, +10 for unprocessed, -5 for processed;

[0124] Defined as a high-weight event, Defined as a medium-weighted event, Defined as a weighted event.

[0125] In a specific embodiment, high-weight information will trigger immediate prompts and subsequent processing actions. The specific weighting algorithm is as follows:

[0126]

[0127] Specifically, the pre-set risk level is quantified based on the inherent risk level of the event. Values ​​that indicate the vehicle deviates from the planned trajectory by more than 0.8 meters, the battery temperature exceeds 75°C, there is abnormal braking, communication is lost, or the distance to an obstacle ahead is less than 3 meters. Values ​​are set to 80-100; when the vehicle deviates from the trajectory by 0.3-0.8m, the battery temperature is 60-75℃, the speed exceeds the limit by ±20%, the image is blurry, or the image acquisition fails. Value set to 50-79; when battery remaining power <25%, ambient temperature >45℃ or <-10℃, and video delay >1s. Set the value to 1-49.

[0128] Duration score It is positively correlated with the duration of time, and the calculation formula is: =5× ,in: For duration in seconds, The upper limit is 100.

[0129] Quantitative presets based on the scope and intensity of the event's impact. Values, events affecting the core systems of the vehicle or areas beyond the vehicle itself, include abnormal braking, battery overheating, sudden approach of obstacles, severe yaw, and loss of communication. Values ​​set to 75-100; events affecting only a single non-core system or localized area include abnormal speed, slight yaw, image blurring, acquisition failure, and high battery temperature. Values ​​are set between 50 and 74; events with minimal impact and no safety association include low battery power, excessively high ambient temperature, excessively low ambient temperature, and video latency. Set the value to 10-49.

[0130] It should be noted that the formulas used to assign different weight levels are weight functions. Weight functions need to comprehensively consider multiple dimensions such as information source, urgency, and reliability, and have the ability to adjust online adaptively.

[0131] The calculation process of the weighting function can be illustrated with an example: four weighting coefficient values ​​are preset, namely... =0.5, =0.2, =0.2, =0.1; When the battery temperature is ≥75℃, =80, duration +5 per second, maximum value 100; initial temperature value =75, exceeding the threshold of 75℃, increment by 1 for every 1℃; not processed. Add 10 to 100; this has been processed. -5 from 0.

[0132] Scenario 1: The battery temperature reaches 85°C and remains untreated for 8 seconds.

[0133] The calculation process is as follows: W = 0.5 × (80 + 10) + 0.2 × (5 × 8) + 0.2 × (75 + 10) + 0.1 × (100 + 10) = 0.5 × 90 + 0.2 × 40 + 0.2 × 85 + 0.1 × 110 = 81 > 80, which is determined to be a high-weight event.

[0134] The present invention provides a remote monitoring method for a large intelligent driving beam transport vehicle.

[0135] Based on the event type and its impact, different weight levels are assigned using a pre-trained feedforward neural network, which includes:

[0136] The input layer is used to input the event type, threshold conditions, duration, and impact level.

[0137] The hidden layer processes the input signal and extracts features, and performs non-linear transformations through an activation function; the activation function used is the Sigmoid function, which compresses the result to between 0 and 1.

[0138] The output layer is used to output the weight scores corresponding to the events.

[0139] In a specific implementation, a feedforward neural network is constructed to learn the mapping relationship between input features and weights for dynamic weight allocation optimization. The input is multi-dimensional features extracted from real-time information streams, and the output is weight levels. It is assumed that each piece of information has multiple features forming a vector. The feedforward neural network is used to calculate the weighted scores of the information. The calculation process is obtained from the following formula:

[0140]

[0141] in:

[0142] : Parameters of the feedforward neural network model;

[0143] : Bias term;

[0144] : A function that compresses the result to a value between 0 and 1, representing the magnitude of the weighted score.

[0145] Furthermore, an error function is established to measure the predicted weight score. With calculated weighted scores The difference between them, the error function is obtained by the following formula:

[0146]

[0147] in:

[0148] Number of samples;

[0149] : No. The calculated weight score is obtained from each piece of information;

[0150] : No. The target weight score is obtained from the prediction of the information.

[0151] The smaller the value, the more accurate the model.

[0152] By constructing a feedforward neural network and establishing an error function for model training, the calculated weight score can be improved. As close as possible to the predicted weight score By continuously adjusting parameters and This makes the calculated weight score similar to the near-predicted weight score. As the error between them decreases, the model can learn how to allocate appropriate weights based on information features.

[0153] The model training process includes:

[0154] The model is divided into training and validation sets for data processing. The training set uses historical event features with manually labeled weights, while the validation set is used for parameter tuning. The model calculates the error by inputting data from the training set, updates the parameters, and iterates the training until the error function converges. The model's accuracy is evaluated using the validation set. New event data is received, and the model parameters are adjusted to adapt to changes in the scenario.

[0155] The remote monitoring method for a large intelligent driving beam transport vehicle provided by this invention issues corresponding operation commands based on real-time weight level information, including:

[0156] When low-weight events are identified:

[0157] For events that are pre-defined as having a minor impact on the safety and performance of the beam transporter and can be handled autonomously by the intelligent driving system of the beam transporter, only recording is performed without triggering active alarms; then the local system of the beam transporter will adjust autonomously; if the low-weight event continues to be displayed for a longer than the preset duration, it will be upgraded to a medium-weight event.

[0158] Specifically, the classification of low-weight events is mainly based on the following criteria: they have a relatively small impact on the operational safety and performance of the beam transport vehicle, and are usually occasional, minor anomalies. Event type quantification... Values ​​range from 1 to 49, duration score Relatively low, the degree of impact is quantified. With a value of 10-49 and a comprehensive weight W < 50, the event has virtually no impact on the core operating function of the beam transporter, and the beam transporter can maintain normal operation for a certain period of time.

[0159] When the remote monitoring module detects a medium-weighted event:

[0160] For events that are pre-set to have a certain impact on the operation safety and performance of the beam transporter, and may be non-core system anomalies that can be handled autonomously by the intelligent driving system of the beam transporter, the system can automatically trigger action responses to reduce the speed and adjust the steering of the beam transporter; specifically, these events may include the following: abnormal speed events and slight yaw events;

[0161] For events that are pre-set to have a certain impact on the safety and performance of the beam transporter, but whose causes are complex and whose intelligent driving system cannot determine the root cause and potential risks, requiring manual review or intervention, they can be displayed in yellow on the monitoring interface, prompting monitoring personnel to choose "ignore," "downgrade," or "upgrade to high weight." Specific events include the following: continuously abnormal speed events, continuously high battery temperature events, continuously blurry images, or events that still fail to acquire data.

[0162] If a medium-weight event continues to be displayed for more than the preset duration, it indicates that the event has not been resolved within the preset time and will be automatically upgraded to a high-weight event. If the yellow indicator for a medium-weight event returns to normal, it will be marked as "resolved" and its weight level will be reduced. Specifically, the classification of medium-weight events is mainly based on the following criteria: they have a certain degree of impact on the operational safety and performance of the beam transport vehicle, and may involve anomalies in some non-core systems of the beam transport vehicle. The event type is quantified. Values ​​range from 50 to 79, duration score Moderate, quantifiable impact The value is 50-74, with a comprehensive weight of 50≤W<80 points. If the event continues to develop, it may have a greater impact on the operation of the beam transport vehicle.

[0163] When the remote monitoring module identifies a high-weight event:

[0164] For events that are pre-set as seriously threatening the operational safety and performance of the beam transporter, possibly involving core system failures or major external risks, but which can be autonomously handled by the beam transporter's intelligent driving system, automatic action responses can be triggered to reduce the speed and adjust the steering of the beam transporter; specifically, these events may include the following: severe vehicle yaw events and abnormal braking events;

[0165] For events that are pre-set as seriously threatening the operational safety and performance of the beam transporter, potentially indicating a core system malfunction or significant external risk, and for which the intelligent driving system of the beam transporter cannot determine the root cause and potential risks, requiring manual review or intervention, a high-priority alert can be issued. This alerts monitoring personnel to promptly assess the situation and issue clear control commands. The beam transporter will then forcibly reduce speed, adjust steering, initiate an emergency stop, or restart the sensing unit based on the pre-set control commands. Specific events that may include the following are also considered: continuous severe yaw events, continuous approaching obstacles ahead, communication interruption events, and continuous battery overheating events.

[0166] Specifically, the classification of high-weight events is mainly based on: events that seriously threaten the operational safety and performance of the beam transporter, events that may involve core system failures of the beam transporter or significant external risks, and events that are quantified by type. Value is 80-100, duration score High, quantifiable impact A score of 75-100 and a comprehensive weight W ≥ 80 indicate that such an event, once it occurs, requires immediate and effective measures to avoid serious accidents.

[0167] Furthermore, the intelligent driving system of the beam transporter autonomously responds to events. After an event occurs, the intelligent driving system of the beam transporter can analyze and judge the event in a short time based on preset algorithms and models, and directly generate corresponding control commands to handle it.

[0168] Events requiring manual review or intervention indicate that the cause of the event is complex, and the intelligent driving system cannot accurately determine the root cause and potential risks of the event solely based on its own algorithm, or that the event continues even after the intelligent driving system of the beam transport vehicle has autonomously responded, requiring manual intervention based on the actual situation.

[0169] In a specific implementation, when a high-weight event is triggered, the remote monitoring module remotely sends control commands to the vehicle end of the beam transporter. The command types include voice prompts, action control, and system maintenance operations. The operation methods include two types: menu quick options and parameterized custom commands, which improves response efficiency and flexibility.

[0170] Furthermore, when a high-weight event is triggered, the remote monitoring module can identify the event and take corresponding actions based on the event type and response strategy. For situations that can be handled autonomously by the intelligent driving system of the beam transport vehicle, such as slight yaw or short-term low pressure, the remote monitoring module will automatically call the local control algorithm to trigger action responses, such as decelerating and adjusting the steering of the beam transport vehicle, without requiring manual intervention. For high-weight events that are preset to require manual review or intervention, such as severe yaw, persistent obstacles ahead, or communication interruption, a high-priority prompt will be displayed on the remote monitoring module interface, prompting monitoring personnel to make timely judgments and issue clear control commands through the operation command module. If the operator does not respond in time, the beam transport vehicle will forcibly decelerate, adjust the steering, initiate an emergency stop, or restart the sensing unit according to the preset commands.

[0171] Furthermore, to ensure timely command response, this application introduces an event priority scheduling mechanism and a multi-threaded asynchronous processing framework, which ensures that the processing of high-priority events is not blocked by low-priority tasks. Simultaneously, a quick response template is set up in the manual command issuance stage, along with preset strategy suggestions, to reduce operation delays; for critical events that have not responded for a long time, the remote monitoring module can automatically trigger emergency default commands to prevent information backlog or mishandling.

[0172] Furthermore, all control commands are synchronized in real time to the vehicle-side controller of the beam transporter via a highly reliable communication link, and the execution status is fed back to the monitoring center, forming a closed loop.

[0173] As a key functional unit in this application, the remote monitoring module undertakes human-computer interaction and remote response control tasks. It is realized by relying on a unified remote monitoring platform. The remote monitoring module integrates an information display interface, a command issuance system, and two-way voice communication function to achieve one-click communication with the on-site driver. It completes low-latency voice interaction through vehicle speakers and microphones, establishing a real-time perception and control closed loop between the monitoring center and the intelligent beam transport vehicle. In addition, this function can also be used for manual confirmation and emergency dispatch in sudden situations or in scenarios with ambiguous recognition, thereby improving remote management and control capabilities.

[0174] The remote monitoring method for large intelligent driving beam transport vehicles provided by this invention combines real-time information streams with events that have already undergone specific operations, and includes changing weight levels and reducing risk warnings, including:

[0175] Once a high-weight event is identified and a response operation is triggered, a closed-loop task node is automatically generated to record the event identifier, initial weight, instruction content, execution feedback, and subsequent tracking status.

[0176] During the operation of the beam transport vehicle, the feedback channel of the beam transport vehicle is continuously monitored, and the weight level of the event is dynamically adjusted according to the changes in the driving status of the beam transport vehicle.

[0177] If a closed-loop task fails to respond within the specified time, its processing priority will be increased, and a reminder will be sent to the monitoring personnel.

[0178] The remote monitoring method for large intelligent driving beam transport vehicles provided by this invention combines weight information that has already undergone specific operations with real-time information flow, and changes the weight level and reduces risk warnings. It also includes: assigning a unique number to each high-weight event; after each high-weight event is identified and triggers a response operation, the relevant control instructions, feedback information, status adjustments and final processing results are all recorded for subsequent retrospective review and analysis.

[0179] Once the remote monitoring module identifies a high-weight event and triggers a response, the monitoring closed-loop processing module automatically generates a closed-loop task node to record the event identifier, initial weight, instruction content, execution feedback, and subsequent tracking status. During operation, the monitoring closed-loop processing module continuously monitors the vehicle feedback channel and dynamically adjusts the event level based on status changes. For example, if the beam transport vehicle's status returns to normal after the instruction is executed, the event weight level can be reduced from "high-weight event" to "medium-weight event" or "low-weight event," and the task node can be closed. If an invalid response is detected, the monitoring closed-loop processing module will extend the monitoring cycle, increase the processing level, or re-initiate the instruction, and prompt monitoring personnel for manual intervention to prevent processing interruption or the omission of critical risks.

[0180] To ensure traceability and consistency throughout the entire process, the monitoring closed-loop processing module assigns a unique number to each high-weight event. Related control commands, feedback information, status adjustments, and final results are all recorded in the processing log database, supporting subsequent retrospective review and analysis. Furthermore, the monitoring closed-loop processing module has a built-in task timeliness management mechanism. If a closed-loop task fails to respond within the specified time, its processing priority will be automatically increased, and a reminder will be sent to monitoring personnel.

[0181] For example, when the remote monitoring module detects that the beam transport vehicle's yaw exceeds the safety threshold, it is initially determined to be a high-weight event. The monitoring terminal immediately issues a "decelerate + voice prompt" command. If the beam transport vehicle's subsequent status feedback indicates that the yaw angle has returned to the normal range, the monitoring closed-loop processing module determines that the task has been completed, the event status is updated to "resolved," and the weight level is simultaneously downgraded. If there is still no effective feedback within the limited time, the monitoring closed-loop processing module will maintain the high-weight mark of the event and issue another pop-up reminder. Through this event-driven processing mechanism, the monitoring closed-loop processing module achieves full-process management from identification, response, feedback to verification, significantly improving the system's autonomous identification and risk response capabilities in complex scenarios, reducing the risk of human error, and enhancing the system's robustness and operational reliability.

[0182] The closed-loop monitoring module is a key technical unit in this application that ensures the integrity of the remote monitoring system's response and operational stability. Its core function is to track and dynamically adjust high-weight events throughout the entire process, ensuring that every critical anomaly can be handled promptly, effectively, and traceably. This module achieves the binding management of instructions issued by the remote interaction module and vehicle execution results by constructing a closed-loop task queue and feedback judgment mechanism.

[0183] The remote monitoring method for a large intelligent driving beam transport vehicle provided by this invention adjusts the weight level of events using the following formula:

[0184]

[0185] in:

[0186] Information item in time Initial weight score at time;

[0187] Operational feedback;

[0188] Real-time status updates;

[0189] : Operational feedback dominance coefficient, indicating the degree of influence of feedback on weight adjustment;

[0190] Real-time status dominant coefficient, indicating the degree of influence of new data on weight adjustment;

[0191] + =1, ensuring the weight update results are normalized. > This indicates that the dominant weight of the processing result has decreased; if the feedback is good, The weighting is reduced; if potential risks still exist, Maintain weight level prompts.

[0192] By setting a formula to adjust the weight level of events, the information items in the response processing can automatically adjust their weight level based on their subsequent status and newly collected data, avoiding information redundancy warnings and false alarms.

[0193] The remote monitoring method for a large intelligent driving beam transport vehicle provided by this invention adjusts the weight level of events and then re-determines whether to continue prompting the corresponding operation instructions or reduce the weight level. The determination process is represented by the following formula:

[0194]

[0195] in:

[0196] Lower the trigger threshold;

[0197] Warning trigger threshold;

[0198] : Positive effect factor of feedback after response;

[0199] : The residual risk factor of the event in the next time step state.

[0200] Specifically, if low-weight events continue to trigger alerts, it will consume monitoring resources and interfere with decision-making for medium- and high-weight events. Even if some events have returned to normal after processing, continuing to display alerts will lead to an accumulation of invalid information and increase manual filtering costs. By setting an alert suppression function, the display logic of alerts can be precisely controlled, ensuring that monitoring resources focus on the risks that truly require attention. The alert suppression function is obtained from the following formula:

[0201]

[0202] The prompt suppression function only displays prompts for medium and high weight events, while automatically suppressing low weight events, focusing on core risks and reducing interference from irrelevant information; when the weight level of an event decreases after processing, the prompts automatically terminate, improving the simplicity of the monitoring interface and the effectiveness of the information; the monitoring interface only presents necessary information, reducing manual screening costs, enabling operators to quickly locate events that need to be responded to, and shortening response time.

[0203] In summary, by fusing multiple types of sensors, this invention achieves comprehensive coverage of the beam transport vehicle's driving status and environmental data, eliminating monitoring blind spots and solving the problem of insufficient information collection dimensions. Through full-coverage real-time information flow collection and a dynamic weighted grading mechanism, combined with the collaborative processing of a rule engine and a feedforward neural network, this invention significantly improves the remote monitoring efficiency and safety of large intelligent driving beam transport vehicles. The dual verification by the rule engine and feedforward neural network improves the accuracy of high-weight event identification, reduces false alarm rates, and achieves effective hierarchical classification and processing of real-time data. The low, medium, and high-level grading processing mechanism shortens the average response time and reduces reliance on manual intervention. Dynamic weight adjustment and task tracking ensure that each risk is traceable and verifiable, enhancing system reliability and improving the closed-loop feedback mechanism for high-weight events. The graphical interface and audio-visual alarm interaction reduce operational complexity and improve monitoring efficiency.

[0204] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention. It should be noted that for those skilled in the art, several improvements and modifications can be made without departing from the technical principles of the present invention, and these improvements and modifications should also be considered within the scope of protection of the present invention.

Claims

1. A remote monitoring method for a large intelligent driving beam transport vehicle, characterized in that, include: Acquire the beam transport vehicle's driving information and form a comprehensive real-time information stream; The real-time information stream with full coverage is marked into three types of events based on weight information: high-weight events, medium-weight events, and low-weight events. Events with high weight are given priority in notification and processing. Issue corresponding operation instructions based on real-time events; For events that have already been processed, combine real-time information flow to change weight levels and reduce risk warnings; The full-coverage real-time information stream is categorized into three types of events based on weight information: high-weight events, medium-weight events, and low-weight events. Set up a rules engine to fully cover the real-time information flow and identify event types and impact levels according to the rules engine; include: Define multi-dimensional rules to form the rule base conditions, which include event type, threshold conditions, duration of time, and degree of impact. Receive a full-coverage real-time information stream; The received full-coverage real-time information stream is compared with the rules base conditions to identify the event type and determine the degree of impact. Different weight levels are assigned based on the type and impact of the event; Based on the event type and its impact, different weight levels are assigned according to the following formula: in: Weighting coefficients; Final weighted score; Event type; Duration score: the longer the duration, the higher the score; Degree of impact; : Feedback processing status, +10 for unprocessed, -5 for processed; Defined as a high-weight event, Defined as a medium-weighted event, Defined as a low-weight event.

2. The remote monitoring method for a large intelligent driving beam transport vehicle according to claim 1, characterized in that, The driving information includes: real-time data of vehicle condition parameters, driving status, and surrounding images.

3. The remote monitoring method for a large intelligent driving beam transport vehicle according to claim 2, characterized in that, The vehicle condition parameters include: the operating status of the power system, braking unit, power module, and suspension system of the beam transport vehicle; Driving status includes: speed, acceleration, angular velocity, heading, and positioning data of the beam transport vehicle; Real-time data of the surrounding images includes images covering the forward, lateral, and blind spot areas of the beam transport vehicle.

4. The remote monitoring method for a large intelligent driving beam transport vehicle according to claim 1, characterized in that, The assignment of different weight levels based on event type and impact is accomplished through a trained feedforward neural network, which includes: The input layer is used to input the event type, threshold conditions, duration, and impact level. The hidden layer processes the input signal and extracts features, and performs non-linear transformations through an activation function; the activation function used is the Sigmoid function, which compresses the result to between 0 and 1. The output layer is used to output the weight scores corresponding to the events.

5. The remote monitoring method for a large intelligent driving beam transport vehicle according to claim 1, characterized in that, Based on real-time events, corresponding operation instructions are issued, including: When low-weight events are identified: The beam transport vehicle responds autonomously, only recording the event without triggering an active alarm; the beam transport vehicle then adjusts autonomously; if a low-weight event continues to be displayed for more than the preset duration, it is upgraded to a medium-weight event. When a medium-weighted event is identified: For events that are pre-set to be handled autonomously by the intelligent driving system of the beam transport vehicle, the system automatically triggers action responses to reduce the speed and adjust the steering of the beam transport vehicle; For events that are pre-defined as having an undetermined root cause and potential risks and require manual review or intervention, the monitoring personnel are prompted to choose to ignore, downgrade, or escalate them to high-weight events. If a medium-weighted event continues to be displayed for a duration exceeding the preset duration, it will be automatically promoted to a high-weighted event. If it returns to normal, it will be marked as resolved and its weight level will be reduced. When high-weight events are identified: For events that are pre-set to be handled autonomously by the intelligent driving system of the beam transport vehicle, the system automatically triggers action responses to reduce the speed and adjust the steering of the beam transport vehicle; For events that are pre-defined as having an undetermined root cause and potential risks, requiring manual review or intervention, a high-priority alert is issued to prompt monitoring personnel to make a timely assessment and issue control commands.

6. The remote monitoring method for a large intelligent driving beam transport vehicle according to claim 1, characterized in that, For events that have already undergone specific actions, combined with a comprehensive real-time information stream, the following adjustments are made to the weighting level and risk reduction alerts: Once a high-weight event is identified and a response operation is triggered, a closed-loop task node is automatically generated to record the event identifier, initial weight, instruction content, execution feedback, and subsequent tracking status. During the operation of the beam transport vehicle, the feedback channel of the beam transport vehicle is continuously monitored, and the weight level of the event is dynamically adjusted according to the changes in the driving status of the beam transport vehicle. If a closed-loop task fails to respond within the specified time, its processing priority will be increased, and a reminder will be sent to the monitoring personnel.

7. The remote monitoring method for a large intelligent driving beam transport vehicle according to claim 6, characterized in that, For events that have already been processed, combined with a full-coverage real-time information flow, the changes in weight levels and risk reduction alerts also include: assigning a unique number to each high-weight event; after each high-weight event is identified and a response operation is triggered, the relevant control instructions, feedback information, status adjustments, and final processing results are all recorded for subsequent retrospective review and analysis.

8. The remote monitoring method for a large intelligent driving beam transport vehicle according to claim 7, characterized in that, The weight level of a change event is obtained using the following formula: in: : The initial weight score of the information item at time t; Operational feedback; Real-time status updates; : Operational feedback dominance coefficient, indicating the degree of influence of feedback on weight adjustment; Real-time status dominant coefficient, indicating the degree of influence of new data on weight adjustment; + =1, ensuring the weight update results are normalized. > This indicates that the dominant weight of the processing result has decreased; if it recovers to within the preset safety threshold range, The weight is reduced; if it does not recover to the preset safety threshold range, Maintain weight level prompts.

9. The remote monitoring method for a large intelligent driving beam transport vehicle according to claim 8, characterized in that, After adjusting the event's weight level, a new determination is made as to whether to continue prompting the corresponding operation command or to reduce the weight level. The determination process is represented by the following formula: in: Lower the trigger threshold; Warning trigger threshold; : Positive effect factor of feedback after response; : The residual risk factor of the event in the next time step state.

10. A remote monitoring system for a large intelligent driving beam transport vehicle, characterized in that, A remote monitoring method for implementing the large intelligent driving beam transport vehicle as described in any one of claims 1-9 includes: The data acquisition module is used to continuously collect driving information during the operation of the beam transport vehicle, including real-time data on vehicle condition parameters, driving status, and surrounding images. The full-coverage real-time information stream generation module is used to integrate the driving information collected by the data acquisition module into the data channel to form a structured and consistent full-coverage real-time information stream. The full-coverage real-time information stream weight training and allocation module is used to automatically label information with different weight levels in the full-coverage real-time information stream of the beam transport vehicle, and to give priority to and process the full-coverage real-time information stream of the beam transport vehicle with higher weight. The remote monitoring module is used to receive the full-coverage real-time information stream from the beam transport vehicle and issue corresponding operation instructions based on the real-time weight level information. The monitoring closed-loop processing module is used to change the weight level and risk warnings based on the weight information that has already undergone specific operations, combined with the full-coverage real-time information flow.

11. The remote monitoring system for the large intelligent driving beam transport vehicle according to claim 10, characterized in that, The remote monitoring module includes an information display interface, a command issuance system, and a two-way voice communication function, which are used to establish a real-time perception and control closed loop with the beam transport vehicle.

12. The remote monitoring system for the large intelligent driving beam transport vehicle according to claim 11, characterized in that, The information display interface adopts a graphical interface, which classifies and presents the real-time information stream of the vehicle according to the preset weight level; high-weight events are highlighted by color coding and visual hierarchy, and accompanied by sound and light alarm prompts, so that monitoring personnel can identify them in time. Each abnormal message includes the trigger time, event type, vehicle number, geographical location, status parameters, and image snapshot. Monitoring personnel can click to view details and select response actions.

Images