Active driving safety robot control system and method based on artificial intelligence

Abstract

The present invention discloses an active driving safety robot control system based on artificial intelligence, which relates to the field of driving safety control systems, including: an alarm level module, which is used to set different types of alarm levels; an alarm type module, which is used to set different types of driving alarms and configure corresponding rules; a processing type module, which monitors the generated alarm data and is used to set different driving safety behavior processing methods; an alarm process module, which is used to set the disposal after the alarm is generated and configure the corresponding execution rules; an AI monitoring robot module, which is the control center for active driving safety behavior monitoring and is used to set the behavior monitoring method; an AI knowledge base module, which is used to set different active safety knowledge bases and push warnings by category. The present invention comprehensively improves the driving safety guarantee level and reduces the accident rate through multi-module collaboration, dividing alarm levels, setting alarm types, introducing intelligent robot processing, building efficient alarm processes and using AI monitoring.

Description

Technical Field

[0001] The present invention relates to the technical field of driving safety control systems, and in particular to an artificial intelligence-based driving active safety robot control system and method. Background Art

[0002] With the rapid development of the transportation industry, the number of vehicles continues to increase, and ensuring driving safety has become a vital task. Traditional driving safety measures mainly rely on the driver's careful driving and the constraints of traffic laws and regulations, but due to the uncertainty of human factors and the complexity of the road environment, traffic accidents still occur from time to time.

[0003] In recent years, the rapid development of artificial intelligence technology has brought new solutions to driving safety. Active safety robot control systems based on artificial intelligence have emerged, aiming to improve vehicle driving safety and reduce accident rates through advanced technical means.

[0004] In the current traffic environment, drivers may face various interferences and risk factors while driving. For example, making phone calls, being distracted, not noticing the approach of vehicles behind, smoking, blocking the camera, and not wearing seat belts can all lead to serious safety hazards. In addition, different road conditions, weather conditions, and vehicle performance will also affect driving safety. Traditional security monitoring systems often have problems such as untimely response, high false alarm rate, and single function, which makes it difficult to meet the high requirements of modern transportation for safety assurance.

[0005] In order to solve these problems, artificial intelligence technology has been introduced into the field of driving safety. Through image recognition, sensor technology and machine learning algorithms, the vehicle's driving status and driver's behavior can be monitored in real time, potential safety risks can be discovered in time, and corresponding alarms and treatment measures can be taken. For example, using cameras and deep learning algorithms, various violations of drivers, such as making phone calls and smoking, can be accurately identified, and alarms can be issued in time to remind the driver to correct them. At the same time, through the analysis of vehicle sensor data, the vehicle's speed, acceleration, steering angle and other parameters can be monitored in real time to determine whether the vehicle is in a safe driving state.

[0006] In addition, with the development of Internet of Things technology, information exchange between vehicles and between vehicles and traffic management systems has become more convenient. The AI-based active safety robot control system can use this information to achieve collaborative monitoring and management of vehicles. For example, when a vehicle detects a dangerous situation on the road ahead, it can promptly transmit the information to surrounding vehicles through the network to remind other vehicles to avoid it. At the same time, the traffic management department can also use the system to understand the driving status of the vehicle in real time, take traffic control measures in time, and improve road traffic efficiency and safety.

[0007] In summary, the AI-based active safety robot control system is an innovative product that meets the needs of the times. It integrates advanced technologies such as AI, sensor technology, and the Internet of Things technology to provide a more reliable and efficient solution for ensuring vehicle driving safety. In the future transportation field, this system is expected to play a more important role and ensure people's travel safety. Summary of the invention

[0008] The present invention proposes an artificial intelligence-based active driving safety robot control system and method to solve the problems mentioned in the above-mentioned prior art.

[0009] In order to achieve the above-mentioned purpose, the present invention adopts the following technical scheme: an artificial intelligence-based driving active safety robot control system, comprising:

[0010] The alarm level module is used to set different types of alarm levels and corresponding responses. The alarm level defines the response level of different alarms and can start different alarm levels based on actual conditions.

[0011] The alarm type module is used to set different driving alarm types and configure corresponding rules. The rule configuration sets the corresponding alarm level and alarm attribute (making or receiving phone calls alarm, distraction alarm, rear approach alarm, smoking alarm, camera obstruction alarm, and no seat belt alarm) according to different alarm types, and sets the corresponding alarm level; different alarm types are recorded as different rule codes, recorded as S={S1, S2, S3…SN}, and then classified according to S={S1, S2, S3…SN}. The rule check judges the integrity of the classified configuration to ensure that the rule is available;

[0012] The processing type module monitors the generated alarm data and is used to set different driving safety behavior processing methods and configure corresponding rules, including alarm data collection, data transmission, and the introduction of AI robots in processing. When processing, the robot receives the alarm information data and obtains the alarm level configuration through the data collection center. It is combined with the corresponding process disposal for automatic process assembly and sent to the data processing center for alarm data processing and execution.

[0013] The alarm process module is used to set how to deal with the alarm after it is generated, configure the corresponding execution rules, and perform dynamic process execution after receiving the data from the processing type module. The process mainly uses natural language processing technology and statistical analysis methods to analyze and process the collected data in real time. The model building unit builds an intention recognition model based on machine learning algorithms and finite element analysis methods to facilitate accurate judgment of the alarm processing process and to respond to the corresponding disposal efficiently. Through the accumulation of data in the field of active driving safety, the algorithm can also continue to iterate, which is convenient for identifying and processing complex alarm processes;

[0014] The AI ​​monitoring robot module is the control center for active driving safety behavior monitoring. It is used to set the behavior monitoring method and conduct corrected monitoring. Through the dynamic input of warning rules, it can identify different images, identify safe behaviors during driving, and provide feedback to different users for timely reminders and alarms. When safe behaviors are detected or preset alarm rules are triggered, SMS notifications are automatically sent for warning information. Alarm rules are set and adjusted according to actual operating conditions.

[0015] The AI ​​knowledge base module is used to set up different active safety knowledge bases and push warnings by category. It can continuously improve based on the existing safety behavior knowledge base. Through the process of acceptance, processing, and feedback, the AI ​​knowledge base can be continuously enriched and improved, which is convenient for better active safety monitoring.

[0016] The group information module manages user group data, identifies user identities and group members in the group, and uses JSON format to customize various group information; formulates a time node sequence set according to the user's personalized push requirements to identify the push plan for the target data service, obtains the user's personalized push plan set, and manages the SMS push of the target data service based on the user's personalized push plan.

[0017] Furthermore, the alarm level module includes the following operation steps:

[0018] A set is set for each alarm level respectively, so as to respond in different levels according to the alarm level, and a set can be formed for the alarm level;

[0019] Set device alarm and platform alarm methods for different alarm levels respectively, realize the set of device alarm method definition and platform alarm method definition under different alarm levels, and be able to combine them, and import the obtained M concentrated information entropy sets into the module as alarm level information sets, and finally obtain alarm level data information.

[0020] Introduce the alarm level response stability formula: ,in Indicates the alarm level response stability, Indicates the total number of alarms. represents the response accuracy of the ith alarm (the number of accurate responses divided by the total number of responses), represents the average response accuracy, represents the maximum response accuracy, Represents the minimum response accuracy. This formula evaluates the stability of the alarm level module response by measuring the fluctuation of the response accuracy and its relationship with the maximum response accuracy and the minimum response accuracy, ensuring that the system can stably respond according to the alarm level under different circumstances.

[0021] Furthermore, the alarm type module includes the following operation steps:

[0022] Used to set different driving alarm types and configure corresponding rules. The rule configuration sets corresponding alarm levels and alarm attributes (making or receiving phone calls alarm, distraction alarm, rear approach alarm, smoking alarm, camera obstruction alarm, and not wearing seat belt alarm) according to different alarm types, and sets corresponding alarm levels;

[0023] Configurations of different alarm types are recorded as different rule codes, recorded as S1, S2, S3...SN, and then classified according to S={S1, S2, S3...SN}. The rule check judges the integrity of the classified configuration to ensure that the rules are available.

[0024] Introduce the alarm type configuration complexity formula: ,in Indicates the complexity of alarm type configuration. Indicates the number of alarm types, Represents the weight coefficient of the ith alarm type (determined by the importance and frequency of the alarm type), Indicates the complexity of the rule configuration of the i-th alarm type (which can be measured by the number of rules and the number of parameters). This formula comprehensively considers the weights of different alarm types and the complexity of rule configuration, and evaluates the configuration complexity of the alarm type module for better optimization and management.

[0025] Furthermore, the processing type module is specifically:

[0026] Set up different driving safety behavior processing methods and configure corresponding rules, including alarm data collection, data transmission, and the introduction of AI robots in the processing. When processing, the robot receives the alarm information data, and obtains the alarm level configuration through the data collection center. It is combined with the corresponding process handling for automatic process assembly, and then sent to the data processing center for alarm data processing and execution.

[0027] Introduce the processing efficiency formula: ,in Indicates the processing efficiency, Indicates the number of correctly handled alarms. Indicates the total number of alarms. Represents the average processing time. This formula comprehensively measures the efficiency of the processing type module by considering the ratio of the number of correct processing times to the total number of alarms and the inverse of the average processing time, ensuring that the system can process alarm data quickly and effectively.

[0028] Furthermore, the steps of the alarm process module are:

[0029] S1.1. Receive alarm push data: Through the JT808 unified protocol, tasks can be sent to the positioning device module. The positioning device module is responsible for receiving and parsing task instructions to ensure the accuracy and completeness of task information;

[0030] S1.2, Emergency alarm push: After the receiving device pushes the alarm, the robot will push the relevant warning and voice information to the relevant person in charge, and push different templates according to different types of personnel;

[0031] S1.3, AI alarm: When the push task is completed, the system urgently sends the emergency alarm voice command to the alarm device through JT808, and detects whether the push alarm is completed. The relevant voice is dynamically recognized according to the alarm information;

[0032] S1.4, high-frequency alarm: The robot sends instructions to the alarm device to capture the active safety behavior correction status and perform identification verification on the server. If there is no correction, the high-frequency alarm process is started and voice commands are continuously sent;

[0033] S1.5, cyclic alarm: detect the behavior after the high-frequency alarm. If it is not corrected, start the cyclic alarm and continue to execute S1.1>S1.2>S1.3>S1.4>S1.5 until the behavior is corrected.

[0034] Introduce the alarm process completeness formula: ,in Indicates the completeness of the alarm process. Indicates the number of alarm process steps that are successfully executed. Indicates the total number of alarm process steps, It represents the ratio of the average execution time of each step to the maximum allowed execution time. This formula evaluates the completeness and execution effect of the alarm process module by considering the ratio of the number of successfully executed steps to the total number of steps and the rationality of the execution time of each step, ensuring that the alarm process can be executed efficiently and completely.

[0035] Furthermore, the AI ​​monitoring robot module specifically includes:

[0036] Data storage unit: stores the data uploaded by the active safety behavior alarm data transmission unit to ensure that the data will not be lost;

[0037] Data analysis unit: processes the transmitted data, uses natural language processing technology and statistical analysis methods to perform real-time analysis and processing of the collected data. The model building unit builds an intention recognition model based on machine learning algorithms and finite element analysis methods, and uses support vector machine SVM and artificial neural network ANN machine learning algorithms to learn and train data information related to active safety behaviors, establish a mapping relationship between intentions and entities, and combine finite element analysis methods to optimize model parameters and improve the accuracy and reliability of the model. The prediction accuracy of the model is not less than 90% on the test set.

[0038] In the AI ​​monitoring robot model construction step, the corresponding captured images are converted into vectors, and the behavior in the images is recognized through a convolutional neural network.

[0039] Convolutional neural networks are usually stacked by convolutional layers, pooling layers, and fully connected layers. The convolutional layer uses multiple different convolution kernels to extract the features of the target and generate feature maps; the pooling layer is used for downsampling, merging the features of adjacent feature maps to reduce the dimension; the fully connected layer plays the role of mapping the learned distributed features to the sample label space.

[0040] Introduce the behavior recognition accuracy improvement formula: ,in Indicates the improvement rate of behavior recognition accuracy, It indicates the accuracy of behavior recognition after optimization. Represents the accuracy of behavior recognition before optimization. This formula is used to measure the improvement of the behavior recognition accuracy of the AI ​​monitoring robot module before and after optimization, ensuring that the system can continuously improve its ability to recognize driving safety behaviors.

[0041] Furthermore, the group information module identifies the user identities and group members in the group, and uses the JSON format to customize multiple group information; formulates a time node sequence set according to the user's personalized push requirements to identify the push plan for the target data service, obtains the user's personalized push plan set, and manages the SMS push of the target data service based on the user's personalized push plan.

[0042] Introduce the formula for group information push accuracy: ,in Indicates the accuracy of group information push. Indicates the number of times information is correctly pushed to the target user. Represents the total number of information pushes. This formula is used to evaluate the accuracy of the group information module's push of user information, ensuring that the system can accurately push relevant information to the target user.

[0043] Furthermore, a driving active safety robot control method based on artificial intelligence includes the following steps:

[0044] Alarm level setting steps: set different types of alarm levels to form alarm level sets, define device alarms and platform alarm modes for different alarm levels, and import alarm level information sets into the system;

[0045] Alarm type configuration steps: set different driving alarm types, configure alarm levels and alarm attribute rules according to alarm types, classify different alarm types and perform rule verification;

[0046] Processing method configuration steps: Set different driving safety behavior processing methods and configure corresponding rules, including alarm data collection, transmission and AI robot process assembly, and send the processed data to the data processing center;

[0047] Alarm process execution steps: Receive alarm push data through JT808 unified protocol, and perform emergency alarm push, AI alarm, high-frequency alarm and cyclic alarm until the safety behavior is corrected;

[0048] AI monitoring steps: Use AI monitoring robot modules to monitor driving safety behaviors, including alarm data storage and analysis, behavior recognition through convolutional neural networks, and the construction of intention recognition models to improve the accuracy of behavior recognition; use batch normalization operations in model construction, including calculating the Jacobian matrix of the input vector and the training sample set, performing denormalization processing, and introducing parameter adjustments;

[0049] Steps to improve the AI ​​knowledge base: Set up different active safety knowledge bases, continuously enrich and improve the AI ​​knowledge base through the acceptance, processing, and feedback processes, and realize classified push warnings;

[0050] Group information management steps: Manage user group data, identify group user identities and members, use JSON format and customize multiple group information, and manage SMS push according to user personalized push requirements.

[0051] The comprehensive effect evaluation formula of the introduction method is as follows: ,in The comprehensive effect evaluation value of the expression method, Indicates the alarm level response stability, Indicates the complexity of alarm type configuration. Indicates the processing efficiency, Indicates the completeness of the alarm process. Indicates the improvement rate of behavior recognition accuracy, Indicates the accuracy of group information push. , , , , , is the weight coefficient, which is set according to the importance of different indicators. This formula comprehensively considers the key indicators of each module and comprehensively evaluates the overall effect of the artificial intelligence-based driving active safety robot control method, so as to continuously optimize and improve the system performance.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] The present invention can finely divide different types of alarm levels and set corresponding response methods through the alarm level module. This allows different measures to be taken according to the severity of various driving safety situations, thereby improving the pertinence and effectiveness of the alarm. At the same time, the combination of device alarms and platform alarms ensures that relevant personnel can receive accurate information in a timely manner and respond quickly.

[0054] The setting of the alarm type module enriches the monitoring content of driving safety. By classifying and configuring rules for a variety of common unsafe behaviors, such as making phone calls and distraction alarms, it can comprehensively detect the driver's bad behavior and greatly reduce the risk of accidents caused by human factors. The rule verification function ensures the integrity and availability of the configuration and improves the reliability of the system.

[0055] The processing type module introduces AI robots to achieve efficient collection, transmission and processing of alarm data. The automatic process assembly function makes the processing process more intelligent and efficient, and improves the response speed and processing capacity of driving safety behaviors.

[0056] The alarm process module uses natural language processing and statistical analysis methods to build an efficient alarm processing process. From receiving alarm push data to emergency alarms, AI alarms, high-frequency alarms and cyclic alarms, it ensures continuous supervision and correction of unsafe behaviors until the behavior is corrected, greatly improving the level of driving safety.

[0057] As the control center, the AI ​​monitoring robot module can accurately identify safe behaviors during driving and provide timely feedback and warnings. Combined with advanced machine learning algorithms and convolutional neural networks, the accuracy of behavior recognition is improved, providing strong technical support for driving safety.

[0058] In summary, the artificial intelligence-based active driving safety robot control system of the present invention, through the synergistic effect of multiple modules, effectively improves the driving safety level, reduces the accident rate, and provides reliable protection for the life and property safety of drivers and passengers. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 A schematic block diagram of the artificial intelligence-based active driving safety robot control system proposed by the present invention;

[0060] Figure 2 A schematic block diagram of the artificial intelligence-based active driving safety robot control method proposed by the present invention;

[0061] Figure 3 This is a schematic diagram of the operation steps of the alarm level module in the driving active safety robot control system based on artificial intelligence of the present invention;

[0062] Figure 4 This is a schematic diagram of the steps of the alarm type module in the driving active safety robot control system based on artificial intelligence of the present invention;

[0063] Figure 5 The figure is a schematic diagram of the steps of the alarm process module in the artificial intelligence-based driving active safety robot control system of the present invention. DETAILED DESCRIPTION

[0064] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0065] In the description of the present invention, it should be understood that the terms "center", "longitudinal", "lateral", "length", "width", "thickness", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside", "outside", "clockwise" and "counterclockwise" indicate orientations or positional relationships based on the orientations or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the referred device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be understood as limiting the present invention.

[0066] In addition, the terms "first" and "second" are used for descriptive purposes only and cannot be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include one or more of the features. In the description of the present invention, the meaning of "multiple" is two or more, unless otherwise clearly and specifically defined. In addition, the terms "installed", "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, and it can be the internal connection of two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances. The present invention will be further described in detail below in conjunction with the accompanying drawings.

[0067] Reference Figure 1-5 : An artificial intelligence-based driving active safety robot control system, comprising:

[0068] The alarm level module is used to set different types of alarm levels and corresponding responses. The alarm level defines the response level of different alarms and can start different alarm levels based on actual conditions.

[0069] The alarm type module is used to set different types of driving alarms and configure corresponding rules. The rule configuration sets the corresponding alarm level and alarm attribute (making or receiving phone calls alarm, distraction alarm, rear approach alarm, smoking alarm, camera obstruction alarm, and no seat belt alarm) according to different alarm types, and sets the corresponding alarm level. Different alarm types are recorded as different rule codes, recorded as S1, S2, S3...SN, and then classified according to S={S1, S2, S3...SN}. The rule check judges the integrity of the classified configuration to ensure that the rules are available.

[0070] The processing type module monitors the generated alarm data and is used to set different driving safety behavior processing methods and configure corresponding rules, including alarm data collection, data transmission, and the introduction of AI robots in processing. When processing, the robot receives the alarm information data and obtains the alarm level configuration through the data collection center. It is combined with the corresponding process disposal for automatic process assembly and sent to the data processing center for alarm data processing and execution.

[0071] The alarm process module is used to set how to deal with the alarm after it is generated, configure the corresponding execution rules, and perform dynamic process execution after receiving the data from the processing type module. The process mainly uses natural language processing technology and statistical analysis methods to analyze and process the collected data in real time. The model building unit builds an intention recognition model based on machine learning algorithms and finite element analysis methods to facilitate accurate judgment of the alarm processing process and to respond to the corresponding disposal efficiently. Through the accumulation of data in the field of active driving safety, the algorithm can also continue to iterate, which is convenient for identifying and processing complex alarm processes;

[0072] The AI ​​monitoring robot module is the control center for active driving safety behavior monitoring. It is used to set the behavior monitoring method and conduct corrected monitoring. Through the dynamic input of warning rules, it can identify different images, identify safe behaviors during driving, and provide feedback to different users for timely reminders and alarms. When safe behaviors are detected or preset alarm rules are triggered, SMS notifications are automatically sent for warning information. Alarm rules are set and adjusted according to actual operating conditions.

[0073] The AI ​​knowledge base module is used to set up different active safety knowledge bases and push warnings by category. It can continuously improve based on the existing safety behavior knowledge base. Through the process of acceptance, processing, and feedback, the AI ​​knowledge base can be continuously enriched and improved, which is convenient for better active safety monitoring.

[0074] The group information module manages user group data, identifies user identities and group members in the group, and uses JSON format to customize various group information; formulates a time node sequence set according to the user's personalized push requirements to identify the push plan for the target data service, obtains the user's personalized push plan set, and manages the SMS push of the target data service based on the user's personalized push plan.

[0075] In the present invention, the alarm level module includes the following operating steps:

[0076] A set is set for each alarm level respectively, so as to respond in different levels according to the alarm level, and a set can be formed for the alarm level;

[0077] Set device alarm and platform alarm methods for different alarm levels respectively, realize the set of device alarm method definition and platform alarm method definition under different alarm levels, and be able to combine them, and import the obtained M concentrated information entropy sets into the module as alarm level information sets, and finally obtain alarm level data information.

[0078] Introduce the alarm level response stability formula: ,in Indicates the alarm level response stability, Indicates the total number of alarms. represents the response accuracy of the ith alarm (the number of accurate responses divided by the total number of responses), represents the average response accuracy, represents the maximum response accuracy, Represents the minimum response accuracy. This formula evaluates the stability of the alarm level module response by measuring the fluctuation of the response accuracy and its relationship with the maximum response accuracy and the minimum response accuracy, ensuring that the system can stably respond according to the alarm level under different circumstances.

[0079] In the present invention, the alarm type module includes the following operating steps:

[0080] Used to set different driving alarm types and configure corresponding rules. The rule configuration sets the corresponding alarm level and alarm attributes (making or receiving phone calls alarm, distraction alarm, rear approach alarm, smoking alarm, camera obstruction alarm, and not wearing seat belt alarm) according to different alarm types, and sets the corresponding alarm level;

[0081] Configurations of different alarm types are recorded as different rule codes, recorded as S={S1, S2, S3…SN}, and then classified according to S={S1, S2, S3…SN}. The rule check makes a completeness judgment on the classified configuration to ensure that the rules are available.

[0082] Introduce the alarm type configuration complexity formula: ,in Indicates the complexity of alarm type configuration. Indicates the number of alarm types, Represents the weight coefficient of the ith alarm type (determined by the importance and frequency of the alarm type), Indicates the complexity of the rule configuration of the i-th alarm type (which can be measured by the number of rules and the number of parameters). This formula comprehensively considers the weights of different alarm types and the complexity of rule configuration, and evaluates the configuration complexity of the alarm type module for better optimization and management.

[0083] In the present invention, the processing type module is specifically:

[0084] Set up different driving safety behavior processing methods and configure corresponding rules, including alarm data collection, data transmission, and the introduction of AI robots in the processing. When processing, the robot receives the alarm information data, and obtains the alarm level configuration through the data collection center. It is combined with the corresponding process handling for automatic process assembly, and then sent to the data processing center for alarm data processing and execution.

[0085] Introduce the processing efficiency formula: ,in Indicates the processing efficiency, Indicates the number of correctly handled alarms. Indicates the total number of alarms. Represents the average processing time. This formula comprehensively measures the efficiency of the processing type module by considering the ratio of the number of correct processing times to the total number of alarms and the inverse of the average processing time, ensuring that the system can process alarm data quickly and effectively.

[0086] In the present invention, the alarm process module steps are:

[0087] S1.1. Receive alarm push data: Through the JT808 unified protocol, tasks can be sent to the positioning device module. The positioning device module is responsible for receiving and parsing task instructions to ensure the accuracy and completeness of task information;

[0088] S1.2, Emergency alarm push: After the receiving device pushes the alarm, the robot will push the relevant warning and voice information to the relevant person in charge, and push different templates according to different types of personnel;

[0089] S1.3, AI alarm: When the push task is completed, the system urgently sends the emergency alarm voice command to the alarm device through JT808, and detects whether the push alarm is completed. The relevant voice is dynamically recognized according to the alarm information;

[0090] S1.4, high-frequency alarm: The robot sends instructions to the alarm device to capture the active safety behavior correction status and perform identification verification on the server. If there is no correction, the high-frequency alarm process is started and voice commands are continuously sent;

[0091] S1.5, cyclic alarm: detect the behavior after the high-frequency alarm. If it is not corrected, start the cyclic alarm and continue to execute S1.1>S1.2>S1.3>S1.4>S1.5 until the behavior is corrected.

[0092] Introduce the alarm process completeness formula: ,in Indicates the completeness of the alarm process. Indicates the number of alarm process steps that are successfully executed. Indicates the total number of alarm process steps, It represents the ratio of the average execution time of each step to the maximum allowed execution time. This formula evaluates the completeness and execution effect of the alarm process module by considering the ratio of the number of successfully executed steps to the total number of steps and the rationality of the execution time of each step, ensuring that the alarm process can be executed efficiently and completely.

[0093] In the present invention, the AI ​​monitoring robot module specifically includes:

[0094] Data storage unit: stores the data uploaded by the active safety behavior alarm data transmission unit to ensure that the data will not be lost;

[0095] Data analysis unit: processes the transmitted data, uses natural language processing technology and statistical analysis methods to perform real-time analysis and processing of the collected data. The model building unit builds an intention recognition model based on machine learning algorithms and finite element analysis methods, and uses support vector machine SVM and artificial neural network ANN machine learning algorithms to learn and train data information related to active safety behaviors, establish a mapping relationship between intentions and entities, and combine finite element analysis methods to optimize model parameters and improve the accuracy and reliability of the model. The prediction accuracy of the model is not less than 90% on the test set.

[0096] In the AI ​​monitoring robot model construction step, the corresponding captured images are converted into vectors, and the behavior in the images is recognized through a convolutional neural network.

[0097] Convolutional neural networks are usually stacked by convolutional layers, pooling layers, and fully connected layers. The convolutional layer uses multiple different convolution kernels to extract the features of the target and generate feature maps; the pooling layer is used for downsampling, merging the features of adjacent feature maps to reduce the dimension; the fully connected layer plays the role of mapping the learned distributed features to the sample label space.

[0098] Introduce the behavior recognition accuracy improvement formula: ,in Indicates the improvement rate of behavior recognition accuracy, It indicates the accuracy of behavior recognition after optimization. Represents the accuracy of behavior recognition before optimization. This formula is used to measure the improvement of the behavior recognition accuracy of the AI ​​monitoring robot module before and after optimization, ensuring that the system can continuously improve its ability to recognize driving safety behaviors.

[0099] In the present invention, the group information module identifies the user identities and group members in the group, and uses the JSON format to customize multiple group information; formulates a time node sequence set according to the user's personalized push requirements to identify the push plan for the target data service, obtains the user's personalized push plan set, and manages the SMS push of the target data service based on the user's personalized push plan.

[0100] Introduce the formula for group information push accuracy: ,in Indicates the accuracy of group information push. Indicates the number of times information is correctly pushed to the target user. Represents the total number of information pushes. This formula is used to evaluate the accuracy of the group information module's push of user information, ensuring that the system can accurately push relevant information to the target user.

[0101] In the present invention, a driving active safety robot control method based on artificial intelligence comprises the following steps:

[0102] Alarm level setting steps: set different types of alarm levels to form alarm level sets, define device alarms and platform alarm modes for different alarm levels, and import alarm level information sets into the system;

[0103] Alarm type configuration steps: set different driving alarm types, configure alarm levels and alarm attribute rules according to alarm types, classify different alarm types and perform rule verification;

[0104] Processing method configuration steps: Set different driving safety behavior processing methods and configure corresponding rules, including alarm data collection, transmission and AI robot process assembly, and send the processed data to the data processing center;

[0105] Alarm process execution steps: Receive alarm push data through JT808 unified protocol, and perform emergency alarm push, AI alarm, high-frequency alarm and cyclic alarm until the safety behavior is corrected;

[0106] AI monitoring steps: Use AI monitoring robot modules to monitor driving safety behaviors, including alarm data storage and analysis, behavior recognition through convolutional neural networks, and the construction of intention recognition models to improve the accuracy of behavior recognition; use batch normalization operations in model construction, including calculating the Jacobian matrix of the input vector and the training sample set, performing denormalization processing, and introducing parameter adjustments;

[0107] Steps to improve the AI ​​knowledge base: Set up different active safety knowledge bases, continuously enrich and improve the AI ​​knowledge base through the acceptance, processing, and feedback processes, and realize classified push warnings;

[0108] Group information management steps: Manage user group data, identify group user identities and members, use JSON format and customize multiple group information, and manage SMS push according to user personalized push requirements.

[0109] The comprehensive effect evaluation formula of the introduction method is as follows: ,in The comprehensive effect evaluation value of the expression method, Indicates the alarm level response stability, Indicates the complexity of alarm type configuration. Indicates the processing efficiency, Indicates the completeness of the alarm process. Indicates the improvement rate of behavior recognition accuracy, Indicates the accuracy of group information push. , , , , , is the weight coefficient, which is set according to the importance of different indicators. This formula comprehensively considers the key indicators of each module and comprehensively evaluates the overall effect of the artificial intelligence-based driving active safety robot control method, so as to continuously optimize and improve the system performance.

[0110] The above are only preferred specific implementation modes of the present invention, but the protection scope of the present invention is not limited thereto. Any technician familiar with the technical field can replace or change the technical solution and inventive concept of the present invention within the technical scope disclosed by the present invention, which should be covered by the protection scope of the present invention.

Claims

1. An artificial intelligence-based active safety robot control system, characterized in that: include: The alarm level module is used to set different types of alarm levels and correspond to different response levels. It also introduces the alarm level response stability formula to determine the fluctuation of the response accuracy and its relationship with the maximum response accuracy and the minimum response accuracy, and evaluates the stability of the alarm level module response. The alarm type module is used to set different driving alarm types and configure corresponding rules. Different alarm type configurations are recorded as different rule codes and classified according to the rule code set. The rule check makes a completeness judgment on the classified configurations. By introducing the alarm type configuration complexity formula, the configuration complexity of the alarm type module is evaluated. The processing type module is used to set the processing methods for different driving safety behaviors, configure the corresponding rules, and introduce AI robots. The AI ​​robots receive alarm information data, obtain the alarm level configuration through the data collection center, and perform automatic process assembly. By introducing the processing efficiency formula, it is used to control and evaluate the efficiency of alarm data; The alarm process module is used to set the handling process after the alarm is generated and configure the corresponding execution rules. The process uses natural language processing technology and statistical analysis methods to analyze and process the collected data in real time. The model building unit builds an intention recognition model based on machine learning algorithms and finite element analysis methods; The AI ​​monitoring robot module includes a data storage unit and a data analysis unit, sets the behavior monitoring method, and performs corrected monitoring, recognizes different images through dynamic input of early warning rules, and builds an AI monitoring robot model; AI knowledge base module, used to set up different active safety knowledge bases and push warnings by category; The group information module manages user group data, identifies user identities and group members in the group, and uses JSON format to customize various group information.

2. The artificial intelligence-based active safety robot control system for driving according to claim 1 is characterized in that: The alarm level module includes the following operating steps: S1. Set a set for each alarm level, so as to respond in different levels according to the alarm level; S2. Set the device alarm and platform alarm modes for different alarm levels respectively, realize the set of device alarm mode definitions and platform alarm mode definitions under different alarm levels, and combine them, import the obtained M concentrated information entropy sets into the module as alarm level information sets, and finally obtain alarm level data information; The alarm level response stability formula is: ,in Indicates the alarm level response stability, Indicates the total number of alarms. represents the response accuracy of the i-th alarm, represents the average response accuracy, represents the maximum response accuracy, Represents the minimum response accuracy.

3. The artificial intelligence-based active safety robot control system for driving according to claim 1 is characterized in that: Different alarm types are configured as different rule codes, which are recorded as S={S1, S2, S3…SN}, and then classified according to S={S1, S2, S3…SN}. The rule checker performs integrity judgment on the classified configuration to ensure that the rule is available. The alarm type module includes: setting the corresponding alarm level by making or receiving phone calls, distraction alarm, rear approach alarm, smoking alarm, probe obstruction alarm, and unfastened seat belt alarm. The alarm type configuration complexity formula is: ,in Indicates the complexity of alarm type configuration. Indicates the number of alarm types, represents the weight coefficient of the i-th alarm type, Indicates the complexity of the rule configuration of the i-th alarm type.

4. The artificial intelligence-based active safety robot control system for driving according to claim 1 is characterized in that: The processing type module introduces the processing efficiency formula: ,in Indicates the processing efficiency, Indicates the number of correctly handled alarms. Indicates the total number of alarms. Indicates the average processing time.

5. The artificial intelligence-based active safety robot control system for driving according to claim 1 is characterized in that: The steps of the alarm process module are: S1.

1. Receive alarm push data: Through the JT808 unified protocol, tasks can be sent to the positioning device module. The positioning device module is responsible for receiving and parsing task instructions to ensure the accuracy and completeness of task information; S1.2, Emergency alarm push: After the receiving device pushes the alarm, the robot will push the relevant warning and voice information to the relevant person in charge, and push different templates according to different types of personnel; S1.3, AI alarm: When the push task is completed, the system urgently sends the emergency alarm voice command to the alarm device through JT808, and detects whether the push alarm is completed. The relevant voice is dynamically recognized according to the alarm information; S1.4, high-frequency alarm: The robot sends instructions to the alarm device to capture the active safety behavior correction status and perform identification verification on the server. If there is no correction, the high-frequency alarm process is started and voice commands are continuously sent; S1.5, cyclic alarm: detect the behavior after the high-frequency alarm. If it is not corrected, start the cyclic alarm and continue to execute S1.1>S1.2>S1.3>S1.4>S1.5 until the behavior is corrected; Introduce the alarm process completeness formula: ,in Indicates the completeness of the alarm process. Indicates the number of alarm process steps that are successfully executed. Indicates the total number of alarm process steps, It represents the ratio of the average execution time of each step to the maximum allowed execution time.

6. The artificial intelligence-based active safety robot control system for driving according to claim 1 is characterized in that: The AI ​​monitoring robot module specifically includes: Data storage unit: stores the data uploaded by the active safety behavior alarm data transmission unit; Data analysis unit: processes the transmitted data, uses natural language processing technology and statistical analysis methods to analyze and process the collected data in real time. The model building unit builds an intention recognition model based on machine learning algorithms and finite element analysis methods, uses support vector machine SVM and artificial neural network ANN machine learning algorithms to learn and train data information related to active safety behaviors, establishes a mapping relationship between intentions and entities, and combines finite element analysis methods to optimize model parameters and improve the accuracy and reliability of the model. The prediction accuracy of the model is not less than 90% on the test set; In the AI ​​monitoring robot model building step, the corresponding captured images are converted into vectors, and the behavior in the images is recognized through a convolutional neural network; Convolutional neural networks are usually stacked by convolutional layers, pooling layers, and fully connected layers. The convolutional layer uses multiple different convolution kernels to extract the features of the target and generate feature maps; the pooling layer is used for downsampling, merging the features of adjacent feature maps to reduce the dimension; the fully connected layer plays the role of mapping the learned distributed features to the sample label space; Introduce the behavior recognition accuracy improvement formula: ,in Indicates the improvement rate of behavior recognition accuracy, It indicates the accuracy of behavior recognition after optimization. Indicates the accuracy of behavior recognition before optimization.

7. The artificial intelligence-based active safety robot control system for driving according to claim 1 is characterized in that: The group information module formulates a time node sequence set according to the user's personalized push requirements to identify the push plan for the target data service, obtains the user's personalized push plan set, and manages the SMS push of the target data service based on the user's personalized push plan, and introduces the group information push accuracy formula: ,in Indicates the accuracy of group information push. Indicates the number of times information is correctly pushed to the target user. Indicates the total number of push messages.

8. A method for applying the artificial intelligence-based active safety robot control system of any one of claims 1 to 7, characterized in that: The following steps are involved: S1.1.1, Alarm level setting steps: Set different types of alarm levels to form an alarm level set, define device alarm and platform alarm methods for different alarm levels, and import the alarm level information set into the system; S1.1.2, Alarm type configuration steps: set different driving alarm types, configure alarm levels and alarm attribute rules according to alarm types, classify different alarm types and perform rule verification; S1.1.3, Processing method configuration steps: Set different driving safety behavior processing methods and configure corresponding rules, including alarm data collection, transmission and AI robot process assembly, and send the processed data to the data processing center; S1.1.4, Alarm process execution steps: Receive alarm push data through JT808 unified protocol, and perform emergency alarm push, AI alarm, high-frequency alarm and cyclic alarm until the safety behavior is corrected; S1.1.5, AI monitoring steps: Use AI monitoring robot modules to monitor driving safety behaviors, including alarm data storage and analysis, behavior recognition through convolutional neural networks, building intention recognition models, and improving the accuracy of behavior recognition; use batch normalization operations in model construction, including calculating the Jacobian matrix of the input vector and the training sample set, performing denormalization processing, and introducing parameter adjustments; S1.1.6, AI knowledge base improvement steps: Set up different active safety knowledge bases, continuously enrich and improve the AI ​​knowledge base through the acceptance, processing, and feedback processes, and realize classified push warnings; S1.1.7, Group information management steps: manage user group data, identify group user identities and members, use JSON format and customize multiple group information, and manage SMS push according to user personalized push requirements; The comprehensive effect evaluation formula of the introduction method is: ,in The comprehensive effect evaluation value of the expression method, Indicates the alarm level response stability, Indicates the complexity of alarm type configuration. Indicates the processing efficiency, Indicates the completeness of the alarm process. Indicates the improvement rate of behavior recognition accuracy, Indicates the accuracy of group information push. , , , , , is the weight coefficient, which is set according to the importance of different indicators.

Images