Piloting method and system based on AI fool-proof technology

Abstract

The invention provides a piloting method and system based on an AI fool-proof technology, and relates to the technical field of intelligent auxiliary driving piloting. Driving environment data, driver state data and current operation behavior sequence data of a driver are obtained and processed in real time; a deep learning model is used for environment perception scoring, driver fatigue degree scoring, driving behavior mode scoring and future driving operation intention prediction, a real-time risk score is calculated based on the scores, a corresponding fool-proof strategy is executed, and the driving safety and reliability can be remarkably improved through the method. Through multi-modal data fusion and a cross-modal attention mechanism, the accuracy and robustness of environment perception are improved, so that the system can make more accurate judgment in a complex and changeable driving environment.

Description

Technical Field

[0001] The present invention relates to the field of intelligent assisted driving and navigation technology, and in particular to a navigation method and system based on AI fool-proofing technology. Background Art

[0002] With the rapid development of the automobile industry and the continuous advancement of intelligent technology, autonomous driving and driver assistance systems have become important means to improve road safety and reduce traffic accidents. Traditional driver assistance systems mainly rely on cameras and lidar for assisted driving navigation, but these systems often have limitations in complex and changeable driving environments and cannot fully and accurately assess driving risks. In recent years, the development of artificial intelligence technology has brought new opportunities for driver assistance systems. By combining multiple sensor data and advanced machine learning algorithms, more accurate environmental perception, driver status monitoring, and behavior pattern analysis can be achieved. However, the existing system still has shortcomings. Although the existing multi-sensor fusion technology can improve the perception accuracy, it lacks an effective dynamic adjustment mechanism in the weight distribution and fusion strategy between different sensor data, and it is difficult to adapt to different driving scenarios. The existing assisted driving navigation process lacks monitoring of the driver's fatigue state. In the assisted driving process, the driver is sometimes required to take over the operation of the vehicle. The degree of fatigue will affect the driver's operation, thereby causing misoperation and causing accidents. Traditional assisted driving navigation does not involve any analysis of driving behavior, or the analysis is usually based on simple rules or statistical models, which makes it difficult to capture complex driving behavior characteristics, has a weak ability to predict future driving operation intentions, and is unable to prepare countermeasures in advance, reducing the system's response speed and safety.

[0003] Therefore, it is necessary to provide a navigation method and system based on AI anti-foolproofing technology to solve the above technical problems. Summary of the invention

[0004] In order to solve the above technical problems, the present invention provides a navigation method and system based on AI anti-foolproofing technology, so as to achieve the beneficial effect of intelligent anti-foolproofing in the process of intelligent assisted driving navigation.

[0005] The present invention provides a navigation method based on AI foolproofing technology, and the navigation method comprises the following steps: S1: Real-time acquisition of driving environment data, driver status data, and driver current operation behavior sequence data; S2: Use the pre-trained deep learning model to process the driving environment data, driver status data, and driver current operation behavior sequence data to obtain the environment perception score, driver fatigue score, driving behavior pattern score, and future driving operation intention prediction; S3: Calculate real-time risk score based on environmental perception score, driver fatigue score and driving behavior pattern score; S4: Execute the error-proofing strategy in real time according to the real-time risk score, and generate the future error-proofing strategy based on the prediction of future driving operation intention and wait for execution.

[0006] Preferably, in step S1, the driver status data includes the driver's facial expression data acquired in real time by a camera device and the driver's eye tracking data acquired in real time by an eye tracking device.

[0007] Preferably, in step S1, the driver's current operation behavior sequence data is a data set of vehicle control actions performed by the driver that are continuously recorded within a preset time window including the current moment.

[0008] Preferably, in step S2, the step of obtaining the environmental perception score is: Acquire driving environment data through different sensors, and perform data preprocessing on the acquired driving environment data; Use a pre-trained convolutional neural network to extract features from the pre-processed driving environment data, and use a cross-modal attention mechanism to dynamically adjust the weights of each sensor data feature to perform multi-modal weighted fusion of data features from different sensors; The data features after multimodal weighted fusion are evaluated through the fully connected layer to obtain the environmental perception score.

[0009] Preferably, in step S2, the step of obtaining the driver fatigue score is: Preprocessing the acquired driver status data; Use pre-trained deep learning models to analyze driver facial expression data and eye tracking data to obtain emotional state scores and attention level scores; The emotional state score and the attention level score are weighted and integrated to obtain the driver fatigue level score.

[0010] Preferably, in step S2, the steps of obtaining the driving behavior pattern score and the future driving operation intention prediction are: Preprocessing the acquired driver's current operation behavior sequence data; Combined with the time series analysis method, key behavior features are extracted from the driver's current operation behavior sequence data; The extracted key behavior features are evaluated using predefined machine learning methods to obtain driving behavior pattern scores, and the extracted key behavior features are processed using long short-term memory networks and Transformer models to obtain future driving operation intention predictions.

[0011] Preferably, the future driving operation intention prediction includes short-term intention prediction, long-term intention prediction and confidence estimation of these two prediction results, wherein the confidence estimation is obtained by quantifying the uncertainty of the results of short-term intention prediction and long-term intention prediction through Bayesian method.

[0012] Preferably, in step S3, the calculation formula of the real-time risk score is: ;

[0013] in, For real-time risk scoring, is the environmental perception score, is the driver fatigue score, is the driving behavior pattern score, , , are their respective dynamic weight coefficients.

[0014] Preferably, each dynamic weight coefficient is dynamically adjusted according to context information through a pre-generated context-aware model.

[0015] The present invention also provides a navigation system based on AI fool-proofing technology, which is applied to a navigation method based on AI fool-proofing technology. The navigation system includes: Data acquisition module, used to obtain driving environment data, driver status data and driver's current operation behavior sequence data in real time; The data analysis module is used to process the driving environment data, driver status data, and driver current operation behavior sequence data using a pre-trained deep learning model to obtain an environmental perception score, a driver fatigue score, a driving behavior pattern score, and a prediction of future driving operation intentions; A real-time risk score calculation module, used to calculate a real-time risk score based on an environmental perception score, a driver fatigue score, and a driving behavior pattern score; The fool-proofing strategy execution module is used to execute the fool-proofing strategy in real time according to the real-time risk score, and to generate future fool-proofing strategies based on the prediction of future driving operation intentions and wait for execution.

[0016] Compared with the related art, the navigation method and system based on AI foolproofing technology provided by the present invention has the following beneficial effects: The present invention acquires and processes driving environment data, driver status data and driver current operation behavior sequence data in real time, uses deep learning model to perform environmental perception scoring, driver fatigue scoring, driving behavior pattern scoring and future driving operation intention prediction, and calculates real-time risk scoring based on these scores, and executes corresponding foolproof strategies. The method can significantly improve the safety and reliability of driving. First, through multimodal data fusion and cross-modal attention mechanism, the accuracy and robustness of environmental perception are improved, so that the system can make more accurate judgments in complex and changeable driving environments; secondly, through the analysis of driver's facial expression and eye tracking data, the emotional state and attention level are comprehensively evaluated, and the comprehensive monitoring of driver fatigue is realized, effectively preventing accidents caused by fatigue driving; thirdly, through time series analysis and machine learning methods, the driving behavior pattern is deeply analyzed and the future driving operation intention is predicted, so that the system can take active measures in advance to reduce potential risks; finally, the weight coefficient is dynamically adjusted through the context perception model, so that the system can flexibly adjust the strategy according to the specific driving environment information, further enhancing the adaptability and response speed of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 This is a flow chart of a navigation method based on AI foolproofing technology of the present invention; Figure 2 The module structure diagram of a navigation system based on AI fool-proofing technology of the present invention. DETAILED DESCRIPTION

[0018] The present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It is to be understood that the specific embodiments described herein are only used to explain the present invention, rather than to limit the present invention. It should also be noted that, for ease of description, only the parts related to the present invention, rather than all structures, are shown in the accompanying drawings. In addition, the embodiments of the present invention and the features in the embodiments may be combined with each other without conflict.

[0019] It should also be noted that, for ease of description, only the parts related to the present invention, but not all of the contents, are shown in the accompanying drawings. Before discussing the exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flow charts. Although the flow charts describe the operations (or steps) as sequential processes, many of the operations therein can be implemented in parallel, concurrently or simultaneously. In addition, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but it can also have additional steps not included in the accompanying drawings. The process can correspond to methods, functions, procedures, subroutines, subprograms, etc.

[0020] Embodiment 1 The present invention provides a navigation method based on AI foolproof technology. In the specific implementation process, Figure 1 As shown, it shows a flowchart of a navigation method based on AI foolproofing technology provided by the present invention, including: Step S1: Acquire driving environment data, driver status data, and driver current operation behavior sequence data in real time.

[0021] Specifically, in step S1, the driver status data includes the driver's facial expression data acquired in real time by a camera device and the driver's eye tracking data acquired in real time by an eye tracking device.

[0022] Specifically, in step S1, the driver's current operation behavior sequence data is a data set of vehicle control actions performed by the driver that are continuously recorded within a preset time window including the current moment.

[0023] During the specific implementation process, the system acquires driving environment data in real time. The driving environment data is collected in real time through a variety of sensors. These sensors are installed at different positions of the vehicle to comprehensively cover the environment around the vehicle. These sensors can provide all-round information about the vehicle's surroundings, including but not limited to road conditions, current vehicle operating status, location and speed of traffic participants, obstacle detection, and driver status data. The driver status data includes driver's facial expression data acquired in real time through a camera device and driver's eye tracking data acquired in real time through an eye tracking device. The driver's current operating behavior sequence data is acquired. The driver's current operating behavior sequence data is a data set of vehicle control actions performed by the driver that are continuously recorded within a preset time window including the current moment. These control actions include but are not limited to steering wheel rotation angle, accelerator pedal position, and brake pedal position. Through these data, the system can fully understand the current navigation status and provide accurate basic information for subsequent analysis and decision-making.

[0024] Step S2: Use the pre-trained deep learning model to process the driving environment data, driver status data, and driver's current operation behavior sequence data to obtain the environment perception score, driver fatigue score, driving behavior pattern score, and future driving operation intention prediction.

[0025] During the specific implementation process, the collected data is processed through a pre-trained deep learning model to obtain relevant scoring results, which are used to comprehensively and accurately evaluate the driving environment and provide a reliable foundation for subsequent risk scoring calculations and fool-proofing strategy execution.

[0026] Specifically, in step S2, the steps for obtaining the environmental perception score are: Acquire driving environment data through different sensors, and perform data preprocessing on the acquired driving environment data; Use a pre-trained convolutional neural network to extract features from the pre-processed driving environment data, and use a cross-modal attention mechanism to dynamically adjust the weights of each sensor data feature to perform multi-modal weighted fusion of data features from different sensors; The data features after multimodal weighted fusion are evaluated through the fully connected layer to obtain the environmental perception score.

[0027] In the specific implementation process, first, the driving environment data is obtained in real time through a variety of sensors installed at different positions of the vehicle. These sensors can provide all-round information about the vehicle's surroundings; next, the acquired driving environment data is preprocessed. The preprocessing steps include data cleaning, denoising, and normalization to ensure the quality and consistency of the input data. The preprocessed data will be used as the input of the convolutional neural network; then, the preprocessed driving environment data is extracted using a pre-trained convolutional neural network. The convolutional neural network performs multi-layer convolution and pooling operations and extracts high-level features. The data from each sensor will be extracted through an independent convolutional neural network branch to generate their own feature representations. In order to further improve the accuracy and robustness of environmental perception, a cross-modal attention mechanism is introduced. The cross-modal attention mechanism can The cross-modal attention mechanism calculates the correlation between the features of each sensor, generates an attention weight matrix, and then performs weighted fusion on the features of each sensor. The system can automatically adjust the importance of different sensor data according to the specific situation of the current driving environment, so as to better cope with complex driving scenarios. Finally, the multi-modal weighted fused data features are evaluated through the fully connected layer to obtain the final environmental perception score. The fully connected layer maps the fused feature vector to a scalar score, which reflects the overall safety and complexity of the current driving environment. The higher the environmental perception score, the more complex the current environment or the higher the risk. Conversely, the lower the score, the simpler and safer the current environment.

[0028] Specifically, in step S2, the steps for obtaining the driver fatigue score are: Preprocessing the acquired driver status data; Use pre-trained deep learning models to analyze driver facial expression data and eye tracking data to obtain emotional state scores and attention level scores; The emotional state score and the attention level score are weighted and integrated to obtain the driver fatigue level score.

[0029] In the specific implementation process, first, the acquired driver status data is preprocessed. The preprocessing steps include data cleaning, denoising, and normalization to ensure the quality and consistency of facial expression data and eye tracking data. These data are collected in real time by cameras and eye tracking devices installed in the cockpit. The camera captures the driver's facial expression changes, including but not limited to blinking frequency, eye opening and closing degree, and facial muscle activity, while the eye tracking device records the driver's eye movement trajectory and gaze point. Next, the preprocessed data is analyzed using a pre-trained deep learning model. Specifically, a convolutional neural network is used to analyze the driver's facial expression data. Feature extraction is performed to identify features related to emotional state and generate an emotional state score. At the same time, the eye tracking data is processed to analyze the driver's eye movement pattern, evaluate the distribution of his attention, and generate an attention level score. Finally, the emotional state score and the attention level score are weighted and fused to obtain the final driver fatigue level score. The weight coefficient can be trained through historical data or preset according to experience. The weighted fusion method can comprehensively consider the driver's emotional state and attention level, so as to more accurately evaluate his fatigue level. The higher the fatigue level score, the more tired the driver is. Conversely, the lower the score, the more awake and alert the driver is.

[0030] Specifically, in step S2, the steps for obtaining the driving behavior pattern score and the future driving operation intention prediction are: Preprocessing the acquired driver's current operation behavior sequence data; Combined with the time series analysis method, key behavior features are extracted from the driver's current operation behavior sequence data; The extracted key behavior features are evaluated using predefined machine learning methods to obtain driving behavior pattern scores, and the extracted key behavior features are processed using long short-term memory networks and Transformer models to obtain future driving operation intention predictions.

[0031] In the specific implementation process, first, the obtained driver's current operation behavior sequence data is preprocessed. The preprocessing steps include data cleaning, normalization, and smoothing to eliminate noise and ensure the consistency and quality of the data; next, combined with the time series analysis method, the key behavior features are extracted from the preprocessed driver's current operation behavior sequence data. The time series analysis method includes but is not limited to statistical-based methods or machine learning-based methods. Through these methods, the system can identify the driver's operation mode and behavior trend, including but not limited to frequent acceleration and deceleration, sharp turns, and maintaining the same direction for a long time; then, the extracted key behavior features are evaluated using a predefined machine learning method to obtain a driving behavior pattern score. The higher the score, the more unstable the current driving behavior or the higher the risk. Conversely, the lower the score, the more stable and safe the driving behavior; at the same time, in order to predict future driving operation intentions, the system uses a long short-term memory network and a Transformer model to process the extracted key behavior features. The long short-term memory network model is good at capturing long-term dependencies and can effectively process complex patterns in time series data. The Transformer model uses the self-attention mechanism to capture the correlation between different time points on a global scale, thereby improving the accuracy of prediction. Through the combination of these two models, the system can predict driving operation intentions in the future.

[0032] Specifically, the prediction of future driving operation intention includes short-term intention prediction, long-term intention prediction and confidence estimation of these two prediction results, wherein the confidence estimation is obtained by quantifying the uncertainty of the results of short-term intention prediction and long-term intention prediction through the Bayesian method.

[0033] In the specific implementation process, the system uses a pre-trained deep learning model to predict future driving operation intentions, including short-term intention predictions, long-term intention predictions, and confidence estimates for these two prediction results. The key behavioral features extracted by the long short-term memory network and the Transformer model are processed to generate short-term intention predictions and long-term intention predictions. In order to quantify the uncertainty of these prediction results, the Bayesian method is used to estimate the confidence of short-term intention predictions and long-term intention predictions. Specifically, the Bayesian method calculates the probability distribution of each prediction result by modeling the posterior probability distribution of the output of the long short-term memory network model and the Transformer model. This posterior probability distribution not only provides the prediction value, but also provides the uncertainty of the prediction value. Through the Bayesian method, the system can generate a confidence estimate for each prediction result, thereby helping the system to better understand and evaluate the reliability of the prediction results and provide a basis for the execution of subsequent fool-proofing strategies.

[0034] Step S3: Calculate a real-time risk score based on the environmental perception score, the driver fatigue score, and the driving behavior pattern score.

[0035] Specifically, in step S3, the calculation formula of the real-time risk score is: ; in, For real-time risk scoring, is the environmental perception score, is the driver fatigue score, is the driving behavior pattern score, , , are their respective dynamic weight coefficients.

[0036] Specifically, each dynamic weight coefficient is dynamically adjusted according to context information through a pre-generated context-aware model.

[0037] In the specific implementation process, the real-time risk score is calculated by weighted fusion of the environmental perception score, the driver fatigue score, and the driving behavior pattern score. In order to ensure the accuracy and adaptability of the real-time risk score, each dynamic weight coefficient is dynamically adjusted according to the current context information through a pre-generated context perception model. The context information includes but is not limited to traffic conditions, weather conditions, road types, time information, and vehicle status. The context perception model automatically adjusts the weight of each scoring item by comprehensively analyzing this information to better reflect the risk level in the current driving environment. For example, in severe weather conditions, the weight of the environmental perception score is may increase, because environmental factors have a greater impact on driving safety at this time; and after long-term driving, the weight of the driver's fatigue score May be increased to pay more attention to the driver's fatigue state, when the driver frequently operates, It may increase to pay more attention to the risk of misoperation caused by the driver's operation. This dynamic adjustment mechanism enables the system to flexibly respond to different driving scenarios and provide more accurate and reliable risk assessment, thereby effectively improving driving safety and reliability.

[0038] Step S4: Execute the foolproof strategy in real time according to the real-time risk score, and generate the future foolproof strategy based on the prediction of future driving operation intention and wait for execution.

[0039] During the specific implementation process, the corresponding fool-proofing strategies are executed in real time according to the real-time risk score, and the future fool-proofing strategies are generated and wait for execution based on the prediction of future driving operation intentions. Specifically, after the system calculates the real-time risk score, it will immediately take corresponding fool-proofing measures according to the preset risk threshold and strategy rules. At the same time, the system uses the prediction of future driving operation intentions, including short-term intention predictions and long-term intention predictions, to generate a series of fool-proofing strategies in advance. These strategies include but are not limited to slowing down in advance, preparing for lane changes, and adjusting the following distance. The generated fool-proofing strategies will be stored and wait for execution. Once the predicted operation intention becomes a reality, the system will immediately execute the corresponding fool-proofing measures, thereby preventing potential risks in advance and ensuring the safety and reliability of driving. Through this combination of real-time response and advance preparation, the system can more effectively respond to various driving scenarios and improve the overall driving experience and safety.

[0040] The working principle of a navigation method based on AI foolproofing technology provided by the present invention is as follows: By acquiring driving environment data, driver status data, including facial expressions and eye tracking data, and driver's current operation behavior sequence data in real time, the pre-trained deep learning model is used for processing to obtain environmental perception scores, driver fatigue scores, driving behavior pattern scores, and future driving operation intention predictions. Based on these scores, real-time risk scores are calculated, and corresponding fool-proofing strategies are executed according to the real-time risk scores. At the same time, future fool-proofing strategies are generated based on future driving operation intention predictions and wait for execution. Among them, the environmental perception score is obtained through multi-sensor data fusion and cross-modal attention mechanism, the driver fatigue score is obtained by analyzing facial expressions and eye tracking data, and the driving behavior pattern score and future driving operation intention prediction are realized through time series analysis and deep learning models. The weight coefficients of each scoring item are dynamically adjusted through the context perception model to adapt to different driving scenarios. This method can comprehensively evaluate driving risks, prevent potential dangers in advance, and improve driving safety and reliability.

[0041] Embodiment 2 The present invention also provides a navigation system based on AI foolproofing technology. In the specific implementation process, Figure 2 As shown, it shows a schematic diagram of the module structure of a navigation system based on AI foolproofing technology provided by the present invention, including: The data acquisition module 100 is used to obtain driving environment data, driver status data and driver current operation behavior sequence data in real time; The data analysis module 200 is used to process the driving environment data, the driver status data, and the driver's current operation behavior sequence data using a pre-trained deep learning model to obtain an environment perception score, a driver fatigue score, a driving behavior pattern score, and a prediction of future driving operation intentions; A real-time risk score calculation module 300, for calculating a real-time risk score based on an environmental perception score, a driver fatigue score, and a driving behavior pattern score; The foolproofing strategy execution module 400 is used to execute the foolproofing strategy in real time according to the real-time risk score, and to generate the future foolproofing strategy according to the prediction of the future driving operation intention and wait for execution.

[0042] The working principle of the navigation system based on AI foolproof technology provided by the present invention is as follows: A navigation system based on AI foolproofing technology includes four main modules: a data acquisition module 100, a data analysis module 200, a real-time risk score calculation module 300 and a foolproofing strategy execution module 400. The data acquisition module 100 acquires driving environment data, driver status data and driver current operation behavior sequence data in real time; the data analysis module 200 processes these data using a pre-trained deep learning model to generate an environmental perception score, a driver fatigue score, a driving behavior pattern score and a future driving operation intention prediction; the real-time risk score calculation module 300 calculates a real-time risk score based on the above scores through a dynamically adjusted weight coefficient; the foolproofing strategy execution module 400 executes the corresponding foolproofing strategy in real time according to the real-time risk score, and generates a future foolproofing strategy based on the future driving operation intention prediction and waits for execution. Through this multi-module collaborative working mode, the system can comprehensively evaluate driving risks and take timely measures to prevent potential dangers, thereby improving driving safety and reliability.

[0043] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0044] A person skilled in the art may understand that all or part of the steps in the various methods of the above embodiments may be completed by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, the storage medium including a read-only memory (ROM), a random access memory (RAM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), a one-time programmable read-only memory (OTPROM), an electronically-erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, magnetic disk storage, magnetic tape storage, or any other computer-readable medium that can be used to carry or store data.

[0045] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

Claims

1. A navigation method based on AI foolproofing technology, characterized in that: The pilot method comprises the following steps: S1: Real-time acquisition of driving environment data, driver status data, and driver current operation behavior sequence data; S2: Use the pre-trained deep learning model to process the driving environment data, driver status data, and driver current operation behavior sequence data to obtain the environment perception score, driver fatigue score, driving behavior pattern score, and future driving operation intention prediction; S3: Calculate real-time risk score based on environmental perception score, driver fatigue score and driving behavior pattern score; S4: Execute the error-proofing strategy in real time according to the real-time risk score, and generate the future error-proofing strategy based on the prediction of future driving operation intention and wait for execution.

2. According to claim 1, a navigation method based on AI foolproof technology is characterized in that: In step S1, the driver status data includes the driver's facial expression data acquired in real time by a camera device and the driver's eye tracking data acquired in real time by an eye tracking device.

3. The navigation method based on AI foolproofing technology according to claim 2 is characterized in that: In step S1, the driver's current operation behavior sequence data is a data set of vehicle control actions performed by the driver that are continuously recorded within a preset time window including the current moment.

4. The navigation method based on AI foolproofing technology according to claim 3 is characterized in that: In step S2, the steps for obtaining the environmental perception score are: Acquire driving environment data through different sensors, and perform data preprocessing on the acquired driving environment data; Use a pre-trained convolutional neural network to extract features from the pre-processed driving environment data, and use a cross-modal attention mechanism to dynamically adjust the weights of each sensor data feature to perform multi-modal weighted fusion of data features from different sensors; The data features after multimodal weighted fusion are evaluated through the fully connected layer to obtain the environmental perception score.

5. The navigation method based on AI foolproofing technology according to claim 4 is characterized in that: In step S2, the steps for obtaining the driver fatigue level score are: Preprocessing the acquired driver status data; Use pre-trained deep learning models to analyze driver facial expression data and eye tracking data to obtain emotional state scores and attention level scores; The emotional state score and the attention level score are weighted and integrated to obtain the driver fatigue level score.

6. The navigation method based on AI foolproofing technology according to claim 5 is characterized in that: In step S2, the steps of obtaining the driving behavior pattern score and the future driving operation intention prediction are: Preprocessing the acquired driver's current operation behavior sequence data; Combined with the time series analysis method, key behavior features are extracted from the driver's current operation behavior sequence data; The extracted key behavior features are evaluated using predefined machine learning methods to obtain driving behavior pattern scores, and the extracted key behavior features are processed using long short-term memory networks and Transformer models to obtain future driving operation intention predictions.

7. The navigation method based on AI foolproofing technology according to claim 6 is characterized in that: The future driving operation intention prediction includes short-term intention prediction, long-term intention prediction and confidence estimation of the two prediction results, wherein the confidence estimation is obtained by quantifying the uncertainty of the results of short-term intention prediction and long-term intention prediction through the Bayesian method.

8. The navigation method based on AI foolproofing technology according to claim 7 is characterized in that: In step S3, the calculation formula of the real-time risk score is: ； in, For real-time risk scoring, is the environmental perception score, is the driver fatigue score, is the driving behavior pattern score, , , are their respective dynamic weight coefficients.

9. The navigation method based on AI foolproofing technology according to claim 8 is characterized in that: Each dynamic weight coefficient is dynamically adjusted according to context information through a pre-generated context-aware model.

10. A navigation system based on AI foolproof technology, characterized in that: Applied to a navigation method based on AI foolproofing technology as claimed in any one of claims 1 to 9, the navigation system comprises: Data acquisition module, used to obtain driving environment data, driver status data and driver's current operation behavior sequence data in real time; The data analysis module is used to process the driving environment data, driver status data, and driver current operation behavior sequence data using a pre-trained deep learning model to obtain an environmental perception score, a driver fatigue score, a driving behavior pattern score, and a prediction of future driving operation intentions; A real-time risk score calculation module, used to calculate a real-time risk score based on an environmental perception score, a driver fatigue score, and a driving behavior pattern score; The fool-proofing strategy execution module is used to execute the fool-proofing strategy in real time according to the real-time risk score, and to generate future fool-proofing strategies based on the prediction of future driving operation intentions and wait for execution.

Images